These sorts of core-density increases are how I win cloud debates in an org.
* Identify the workloads that haven't scaled in a year. Your ERPs, your HRIS, your dev/stage/test environments, DBs, Microsoft estate, core infrastructure, etc. (EDIT, from zbentley: also identify any cross-system processing where data will transfer from the cloud back to your private estate to be excluded, so you don't get murdered with egress charges)
* Run the cost analysis of reserved instances in AWS/Azure/GCP for those workloads over three years
* Do the same for one of these high-core "pizza boxes", but amortized over seven years
* Realize the savings to be had moving "fixed infra" back on-premises or into a colo versus sticking with a public cloud provider
Seriously, what took a full rack or two of 2U dual-socket servers just a decade ago can be replaced with three 2U boxes with full HA/clustering. It's insane.
Back in the late '10s, I made a case to my org at the time that a global hypervisor hardware refresh and accompanying VMware licenses would have an ROI of 2.5yrs versus comparable AWS infrastructure, even assuming a 50% YoY rate of license inflation (this was pre-Broadcom; nowadays, I'd be eyeballing Nutanix, Virtuozzo, Apache Cloudstack, or yes, even Proxmox, assuming we weren't already a Microsoft shop w/ Hyper-V) - and give us an additional 20% headroom to boot. The only thing giving me pause on that argument today is the current RAM/NAND shortage, but even that's (hopefully) temporary - and doesn't hurt the orgs who built around a longer timeline with the option for an additional support runway (like the three-year extended support contracts available through VARs).
If we can't bill a customer for it, and it's not scaling regularly, then it shouldn't be in the public cloud. That's my take, anyway. It sucks the wind from the sails of folks gung-ho on the "fringe benefits" of public cloud spend (box seats, junkets, conference tickets, etc...), but the finance teams tend to love such clear numbers.
The main cost with on-prem is not the price of the gear but the price of acquiring talent to manage the gear. Most companies simply don't have the skillset internally to properly manage these servers, or even the internal talent to know whether they are hiring a good infrastructure engineer or not during the interview process.
For those that do, your scaling example works against you. If today you can merge three services into one, then why do you need full time infrastructure staff to manage so few servers? And remember, you want 24/7 monitoring, replication for disaster recovery, etc. Most businesses do not have IT infrastructure as a core skill or differentiator, and so they want to farm it out.
> even the internal talent to know whether they are hiring a good infrastructure engineer or not during the interview process.
This is really the core problem. Every time I’ve done the math on a sizable cloud vs on-prem deployment, there is so much money left on the table that the orgs can afford to pay FAANG-level salaries for several good SREs but never have we been able to find people to fill the roles or even know if we had found them.
The numbers are so much worse now with GPUs. The cost of reserved instances (let alone on-demand) for an 8x H100 pod even with NVIDIA Enterprise licenses included leaves tens of thousands per pod for the salary of employees managing it. Assuming one SREs can manage at least four racks the hardware pays for itself, if you can find even a single qualified person.
This factually did not play out like this in my experience.
The company did need the same exact people to manage AWS anyway. And the cost difference was so high that it was possible to hire 5 more people which wasn't needed anyway.
Not only the cost but not needing to worry about going over the bandwidth limit and having soo much extra compute power made a very big difference.
Imo the cloud stuff is just too full of itself if you are trying to solve a problem that requires compute like hosting databases or similar. Just renting a machine from a provider like Hetzner and starting from there is the best option by far.
> The main cost with on-prem is not the price of the gear but the price of acquiring talent to manage the gear. Most companies simply don't have the skillset internally to properly manage these servers, or even the internal talent to know whether they are hiring a good infrastructure engineer or not during the interview process.
That's partially true; managing cloud also takes skill, most people forget that with end result being "well we saved on hiring sysadmins, but had to have more devops guys". Hell I manage mostly physical infrastructure (few racks, few hundred VMs) and good 80% of my work is completely unrelated to that, it's just the devops gluing stuff together and helping developers to set their stuff up, which isn't all that different than it would be in cloud.
> And remember, you want 24/7 monitoring, replication for disaster recovery, etc.
And remember, you need that for cloud too. Plenty of cloud disaster stories to see where they copy pasted some tutorial thinking that's enough then surprise.
There is also partial way of just getting some dedicated servers from say OVH and run infra on that, you cut out a bit of the hardware management from skillset and you don't have the CAPEX to deal with.
But yes, if it is less than at least a rack, it's probably not worth looking for onprem unless you have really specific use case that is much cheaper there (I mean less than usual half)
You need the exact same people to run the infra in the cloud. If they don't have IT at all, they aren't spinning up cloud VMs. You're mixing together SaaS and actual cloud infra.
> main cost with on-prem is not the price of the gear but the price of acquiring talent to manage the gear
Not quite. If you hire a bad talent to manage your 'cloud gear' then you would find what the mistakes which would cost you nothing on-premises would cost you in the cloud. Sometimes - a lot.
Is it still a problem in 2026 when unemployment in IT is rising? Reasons can be argued (the end of ZIRP or AI) but hiring should be easier than it was at any time during the last 10 years.
> The main cost with on-prem is not the price of the gear but the price of acquiring talent to manage the gear. Most companies simply don't have the skillset internally to properly manage these servers
This comes up again and again. It was the original sales pitch from cloud vendors.
Often the very same companies repeating this messaging are recruiting and paying large teams of platform developers to manage their cloud…and pay for them to be on call.
While I agree with you, some solutions, such as Oxide Computing could come pretty close to having all the ease of cloud, one whole rack of computers at a time.
To me this doesn't sound logical because you still have to hire someone to manage your cloud deployments which is an entire specialized discipline. Yeah you can get some leeway the job being fully remote I guess but ultimately you aren't reducing headcount as linearly as you seem to imply by going cloud vs on-prem.
Except they do have IT infra as a skill - or did - and replaced it with higher-paid “cloud architects” who just manage VM fleets and DBaaS in the vendor they’re certified in, which could just as easily be done on-site nowadays. They’re lower-skilled overall but more specialized, and thus command higher pay.
Generalists (it me) typically command lower market rates, remain far more flexible, and can replicate much of the experience on-prem while knowing when and why to put something in the public cloud. A few examples:
* 24/7 Telemetry and Monitoring: if you have the talent to roll your own with OpenTelemetry, Grafana, Prometheus, and a database to store the telemetry involved, then great! If it’s me wrangling a hybrid environment though, I’m likely leaning on New Relic to save on headcount and deliver similar results.
* DBaaS: this is increasingly just offered by hypervisor managers since it’s often just spooling up a container and pointing to storage. Not quite a “solved problem”, but enough of one that a single DBA can cover both estates if needed - or a moderately skilled generalist can at least secure it for internal use before offering it out to customers
* Vulnerability Management: as much as I’d love to have an internal Red Team, it’s an order of magnitude cheaper to leverage Nessus, Wazuh, Wiz, or any of the other fleets of continuous scanners to identify vulnerabilities or misconfigurations
That’s just me cherry-picking. The point is less “do everything possible on-prem”, and more “diversify workload placement depending on cost advantages relative to risk models”, and on-prem wins quite handily for a lot of LOB software that just quietly sits and does its thing with minimal fuss.
Given how good Apple Silicon is these days, why not just buy a spec'd out Mac Studio (or a few) for $15k (512 GB RAM, 8 TB NVMe), maybe pay for S3 only to sync data across machines. No talent required to manage the gear. AWS EC2 costs for similar hardware would net out in something ridiculous like 4 months.
That’s definitely the right call in some cases. But as soon as there’s any high-interconnect-rate system that has to be in cloud (appliances with locked in cloud billing contracts, compute that does need to elastically scale and talks to your DB’s pizza box, edge/CDN/cache services with lots of fallthrough to sources of truth on-prem), the cloud bandwidth costs start to kill you.
I’ve had success with this approach by keeping it to only the business process management stacks (CRMs, AD, and so on—examples just like the ones you listed). But as soon as there’s any need for bridging cloud/onprem for any data rate beyond “cronned sync” or “metadata only”, it starts to hurt a lot sooner than you’d expect, I’ve found.
Yep, 100%, but that's why identifying compatible workloads first is key. A lot of orgs skip right to the savings pitch, ignorant of how their applications communicate with one another - and you hit the nail on the head that applications doing even some processing in a cloud provider will murder you on egress fees by trying to hybrid your app across them.
Folks wanting one or the other miss savings had by effectively leveraging both.
What has surprised me about the cloud is that the price has been towards ever increasing prices for cores. Yet the market direction is the opposite, what used to be a 1/2 or a 1/4 of a box is now 1/256 and its faster and yet the price on the cloud has gone ever up for that core. I think their business plan is to wipe out all the people who used to maintain the on premise machines and then they can continue to charge similar prices for something that is only getting cheaper.
Its hard drive and SSD space prices that stagger me on the cloud. Where one of the server CPUs might only be about 2x the price of buy a CPU for a few years if you buy less in a small system (all be it with less clock speed usually on the cloud) the drive space is at least 10-100x the price of doing it locally. Its got a bit more potential redudency but for that overhead you can repeat that data a lot of times.
As time has gone on the deal of cloud has got worse as the hardware got more cores.
Do note though that AIUI these are all E-cores, have poor single-threaded performance and won't support things like AVX512. That is going to skew your performance testing a lot. Some workloads will be fine, but for many users that are actually USING the hardware they buy this is likely to be a problem.
If that's you then the GraniteRapids AP platform that launched previously to this can hit similar numbers of threads (256 for the 6980P). There are a couple of caveats to this though - firstly that there are "only" 128 physical cores and if you're using VMs you probably don't want to share a physical core across VMs, secondly that it has a 500W TDP and retails north of $17000, if you can even find one for sale.
Overall once you're really comparing like to like, especially when you start trying to have 100+GbE networking and so on, it gets a lot harder to beat cloud providers - yes they have a nice fat markup but they're also paying a lot less for the hardware than you will be.
Most of the time when I see takes like this it's because the org has all these fast, modern CPUs for applications that get barely any real load, and the machines are mostly sitting idle on networks that can never handle 1/100th of the traffic the machine is capable of delivering. Solving that is largely a non-technical problem not a "cloud is bad" problem.
These Intel Darkmont cores are in a different performance class than the (Crestmont) E-cores used in the previous generation of Sierra Forest Xeon CPUs. For certain workloads they may have even a close to double performance per core.
Darkmont is a slightly improved variant of the Skymont cores used in Arrow Lake/Lunar Lake and it has a performance very similar to the Arm Neoverse V3 cores used in Graviton5, the latest generation of custom AWS CPUs.
However, a Clearwater Forest Xeon CPU has much more cores per socket than Graviton5 and it also supports dual-socket motherboards.
Darkmont also has a greater performance than the older big Intel cores, like all Skylake derivatives, inclusive for AVX-using programs, so it is no longer comparable with the Atom series of cores from which it has evolved.
Darkmont is not competitive in absolute performance with AMD Zen 5, but for the programs that do not use AVX-512 it has better performance per watt.
However, since AMD has started to offer AVX-512 for the masses, the number of programs that have been updated to be able to benefit from AVX-512 is increasing steadily, and among them are also applications where it was not obvious that using array operations may enhance performance.
Because of this pressure from AMD, it seems that this Clearwater Forest Xeon is the final product from Intel that does not support AVX-512. Both next 2 Intel CPUs support AVX-512, i.e. the Diamond Rapids Xeon, which might be launched before the end of the year, and the desktop and laptop CPU Nova Lake, whose launch has been delayed to next year (together with the desktop Zen 6, presumably due to the shortage of memories and production allocations at TSMC).
E-cores aren't that slow, yesteryear ones were already around Skylake levels of performance (clock for clock). Now one might say that's a 10+ year old uarch, true, but those ten years were the slowest ten years in computing since the beginning of computing, at least as far as sequential programs are concerned.
At my job we use HyperV, and finding someone who actually knows HyperV is difficult and expensive. Throw in Cisco networking, storage appliances, etc to make it 99.99% uptime...
Also that means you have just one person, you need at least two if you don't want gaps in staffing, more likely three.
Then you still need all the cloud folks to run that.
We have a hybrid setup like this, and you do get a bit of best of both worlds, but ultimately managing onprem or colo infra is a huge pain in the ass. We only do it due to our business environment.
I think you're hitting on a general problem statement a lot of orgs run into, even ignoring the uptime figure...
All of the complexity of onprem, especially when you need to worry about failover/etc can get tricky, especially if you are in a wintel env like a lot of shops are.
i.e. lots of companies are doing sloppy 'just move the box to an EC2 instance' migrations because of how VMWare jacked their pricing up, and now suddenly EC2/EBS/etc costing is so cheap it's a no brain choice.
I think the knowledge base to set up a minimal cost solution is too tricky to find a benefit vs all the layers (as you almost touched on, all the licensing at every layer vs a cloud provider managing...)
That said, rug pulls are still a risk; I try to push for 'agnostic' workloads in architecture, if nothing else because I've seen too many cases where SaaS/PaaS/etc decide to jack up the price of a service that was cheap, and sure you could have done your own thing agnostically, but now you're there, and migrating away has a new cost.
IOW, I agree; I don't think the human capital is there as far as infra folks who know how to properly set up such environments, especially hitting the 'secure+productive' side of the triangle.
> These sorts of core-density increases are how I win cloud debates in an org.
AMD has had these sorts of densities available for a minute.
> Identify the workloads that haven't scaled in a year.
I have done this math recently, and you need to stop cherry picking and move everything. And build a redundant data center to boot.
Compute is NOT the major issue for this sort of move:
Switching and bandwidth will be major costs. 400gb is a minimum for interconnects and for most orgs you are going to need at least that much bandwidth top of rack.
Storage remains problematic. You might be able to amortize compute over this time scale, but not storage. 5 years would be pushing it (depending on use). And data center storage at scale was expensive before the recent price spike. Spinning rust is viable for some tasks (backup) but will not cut it for others.
Human capital: Figuring out how to support the hardware you own is going to be far more expensive than you think. You need to expect failures and staff accordingly, that means resources who are going to be, for the most part, idle.
Cloud = the right choice when just starting. It isn't about infra cost, it is about mental cost. Setting up infra is just another thing that hurts velocity. By the time you are serving a real load for the first time though you need to have the discussion about a longer term strategy and these points are valid as part of that discussion.
I guess it depends, but infra is also a lot simpler when starting out. It really isnt much harder (easier even?) to setup services on a box or two than managing AWS.
Im pretty sure a box like this could run our whole startup, hosting PG, k8s, our backend apis, etc, would be way easier to setup, and not cost 2 devops and $40,000 a month to do it.
Is infra really that hard to set up? It seems like infra is something a infra expert could establish to get the infra going and then your infra would be set up and you would always have infra.
It seems a lot of people have forgotten how BigCorp IT used to work.
- request some HW to run $service
- the "IT dept" (really, self-interested gatekeeper) might give you something now, or in two weeks, or god help you if they need to order new hardware then its in two months, best case
- there will be various weird rules on how the on-prem HW is run, who has access etc, hindering developer productivity even further
- the hardware might get insanely oversubscribed so your service gets half a cpu core with 1GB RAM, because perverse incentives mean the "IT dept" gets rewarded for minimizing cost, while the price is paid by someone else
- and so on...
The cloud is a way around this political minefield.
Is that personnel cost more than running on someone else's infra? Just counting the amount of people a company now need just to maintain their cloud/kubernetes/whatever setup, paired with "devops" meaning all devs now have to spend time on this stuff, I could almost wager we would spend less on personnel if we just chucked a few laptops in a closet and sshed in.
Is using virtualization the only good way of taking a 288-core box and splitting it up into multiple parallel workloads? One time I rented a 384-core AMD EPYC baremetal VM in GCP and I could not for the life of me get parallelized workloads to scale just using baremetal linux. I wanted to run a bunch of CPU inference jobs in parallel (with each one getting 16 cores), but the scaling was atrocious - the more parallel jobs you tried to add, the slower all of them ran. When I checked htop the CPU was very underutilized, so my theory was that there was a memory bottleneck somewhere happening with ONNX/torch (something to do with NUMA nodes?) Anyway, I wasn't able to test using proxmox or vmware on there to split up cpu/memory resources; we decided instead to just buy a bunch of smaller-core-count AMD Ryzen 1Us instead, which scaled way better with my naive approach.
They are used for VMs because the load is pretty spiky and usually not that memory heavy. For just running single app smaller core count but higher clocked ones are usually more optimal
>Anyway, I wasn't able to test using proxmox or vmware on there to split up cpu/memory resources; we decided instead to just buy a bunch of smaller-core-count AMD Ryzen 1Us instead, which scaled way better with my naive approac
If that was single 384 (192 times 2 for hyperthreading) CPU you are getting "only" 12 DDR5 channels, so one RAM channel is shared by 16c/32y
So just plain 16 core desktop Ryzen will have double memory bandwidth per core
How did the speed of one or two jobs on the EPYC compare to the Ryzen?
And 384 actual cores or 384 hyperthreading cores?
Inference is so memory bandwidth heavy that my expectations are low. An EPYC getting 12 memory channels instead of 2 only goes so far when it has 24x as many cores.
> These sorts of core-density increases are how I win cloud debates in an org.
The core density is bullshit when each core is so slow that it can't do any meaningful work. The reality is that Intel is 3 times behind AMD/TSMC on performance vs power consumption ratio.
People would be better off having a look at the high frequency models (9xx5F models like the 9575F), that was the first generation of CPU server to reach ~5 GHz and sustain it on 32+ cores.
Intel seem to be deliberately hiding the clock frequency of this thing, the xeon-6-plus-product-deck.pdf has no mention of clock frequency or how LLC is shared.
> If we can't bill a customer for it, and it's not scaling regularly, then it shouldn't be in the public cloud. That's my take, anyway. It sucks the wind from the sails of folks gung-ho on the "fringe benefits" of public cloud spend (box seats, junkets, conference tickets, etc...), but the finance teams tend to love such clear numbers.
I agree, but.
For one, it's not just the machines themselves. You also need to budget in power, cooling, space, the cost of providing redundant connectivity and side gear (e.g. routers, firewalls, UPS).
Then, you need a second site, no matter what. At least for backups, ideally as a full failover. Either your second site is some sort of cloud, which can be a PITA to set up without introducing security risks, or a second physical site, which means double the expenses.
If you're a publicly listed company, or live in jurisdictions like Europe, or you want to have cybersecurity insurance, you have data retention, GDPR, SOX and a whole bunch of other compliance to worry about as well. Sure, you can do that on-prem, but you'll have a much harder time explaining to auditors how your system works when it's a bunch of on-prem stuff vs. "here's our AWS Backup plans covering all servers and other data sources, here is the immutability stuff, here are plans how we prevent backup expiry aka legal hold".
Then, all of that needs to be maintained, which means additional staff on payroll, if you own the stuff outright your finance team will whine about depreciation and capex, and you need to have vendors on support contracts just to get firmware updates and timely exchanges for hardware under warranty.
Long story short, as much as I prefer on-prem hardware vs the cloud, particularly given current political tensions - unless you are a 200+ employee shop, the overhead associated with on-prem infrastructure isn't worth it.
> Then, you need a second site, no matter what. At least for backups, ideally as a full failover. Either your second site is some sort of cloud, which can be a PITA to set up without introducing security risks, or a second physical site, which means double the expenses.
You can technically have backblaze's unlimited backup option which costs around 7$ for a given machine although its more intended for windows, there have been people who make it work and Daily backups and it should work with gdpr (https://www.backblaze.com/company/policy/gdpr) with something like hetzner perhaps if you are worried about gdpr too much and OVH storage boxes (36 TB iirc for ~55$ is a good backup box) and you should try to follow 3-2-1 strategy.
> Then, all of that needs to be maintained, which means additional staff on payroll, if you own the stuff outright your finance team will whine about depreciation and capex, and you need to have vendors on support contracts just to get firmware updates and timely exchanges for hardware under warranty.
I can't speak for certain but its absolutely possible to have something but iirc for companies like dell, its possible to have products be available on a monthly basis available too and you can simply colocate into a decent datacenter. Plus points in that now you can get 10-50 GB ports as well if you are too bandwidth hungry and are available for a lot lot more customizable and the hardware is already pretty nice as GP observed. (Yes Ram prices are high, lets hope that is temporary as GP noted too)
I can't speak about firmware updates or timely exchanges for hardware under security.
That being said, I am not saying this is for everyone as well. It does essentially boils down to if they have expertise in this field/can get expertise in this field or not for cheaper than their aws bills or not. With many large AWS bills being in 10's of thousands of dollars if not hundreds of thousands of dollars, I think that far more companies might be better off with the above strategy than AWS actually.
>Realize the savings to be had moving "fixed infra" back on-premises or into a colo versus sticking with a public cloud provider
As other people have pointed out: what happens when the PSU or mobo shits itself, what happens when the new version of vmware or docker (or whatever) shits itself, etc
I don't post much on HN, but this topic is near and dear to my heart, so here we go.
Context: been helpdesk, sysadmin & network admin, DevOps, Site Reliability Engineer, in that progression, starting in the 90's. Max on-prem was 40 racks, scaled up and down over years.
Many comments talking about how staffing is a key element of this equation that can't be overlooked, but I decided to reply to the root comment, which doesn't say whether/how it considers staffing.
This is a complex equation - and it is relatively easy to present an incomplete or misleading picture management to push the move into the cloud.. or out of the cloud.
Some factors, in no particular order:
1) Scaling: it is self-evident that pulling a single physical server worth out of the cloud is not worth it.. even for 288 cores. Or perhaps 1152 for 4xXeon in a single server. Still likely not worth it. Why? Because a single server is never just that. Someone has to swap components when it goes down. When it goes down.. ALL 1152 cores are down, along with everything they are doing. Is that acceptable for all applications running on all those cores? It is also appropriate supporting infrastructure - power, cooling, physical space. The "fairly obvious" minimum scaling is "enough servers that one can be entirely down for maintenance while keeping everything else running." But now you're paying for some overhead. At 2 servers, you're buying 2x what you need, half that capacity is idle all the time. And so on.
On this point - I think the other comments talking about "each SRE managing 4 (or 5, or 7)" racks missed the point entirely. SRE's should be doing scalable work, whether in the cloud, or on-prem. And they should NOT be swapping failed hard drives and power supplies. Designing a larger-than-one rack install is probably worth hiring consultants for if you don't have that expertise in-house, though the SREs that would be supporting it would need to supply lots of input. To some extent, server & network equipment vendors can also help. It is not trivial as the scale goes up. But then it should run for some years, with relatively unskilled people handling hardware failures and you can re-engage consultants if necessary to do upgrades as hardware and needs evolve.
But your SREs should be on-staff, and probably on-call to handle the software running on that hardware.. and to some extent to call the remote hands to deal with hardware failures.
2) Business needs: does the business need the tech skills that self-hosting requires for the core business? For example - if the business itself is cloud SAAS, maybe DIY-ing at least some of your infrastructure is right in your wheelhouse. If so - a modest increase in staff could mean a huge cost savings. But if not, all the cost of skilled staff to run it is simply part of the cost of in-housing this stuff.
3) Staffing: the people that swap broken hardware are not the same people that respond to pages because the business critical application crashed due to a bug. You can pay a colo facility for all this, typically by the hour - but it isn't cheap and you've got to supply all the spares etc. Is that part of your budget for on-prem?
4) On-call: maybe your self-hosted ERP system can be down every night and weekend without issues.. and even business hours can tolerate 98% uptime. But that doesn't mean you can get away without having someone on call - presumably you're hosting more than just this lowish-requirement ERP system. I'll disagree with other comments - the "no burnout" number of on-call staff you need is 6-7, not 4! Remember people take vacations too. This is well studied, and established, I'll reference Tom Limoncelli's books. This could be relatively cheap and require fewer staff with geo-distributed staff, and it would tend to overlap with staff you already use to provide on-call for anything you host on the cloud - so maybe for your situation it is close to a wash. But you can't forget to budget for it even if the line item is $0.
5) Vendor support: maybe you already have your own data center or colo and are hosting a ton of stuff. Why not move all your Atlassian stuff in house and save the hosting cost. Oh.. wups, Atlassian simply doesn't support that any more. Host it with Atlassian or GTFO. A minor point as most vendors would give you enough notice you can simply run out the lifetime of the hardware it is on and not replace it.
6) Market pricing: At one point Amazon was starting "by the minute, by the core cloud" (as opposed the older "cloudish" model of leasing only an entire physical server by the entire year) and priced a bit under market to get going. Then once they established dominance, they cranked the price WAY up for profit extraction. But now they do have some competition and they're a bit more selective about how they extract profits. In my perception they've shifted a lot of the profit taking to the value-added services rather than raw instance time, but I could be wrong. And they have HUGE costs - it is beyond naive to look at the per-hour cost of an instance and compare it to purchasing an identical physical server solely on the purchase price of that server. Corey Quinn / Duckbill group has spent a huge part of his career in this space - if you're already in the cloud 100% it is well worth optimizing those costs before you start comparing it to what on-prem might cost.
> The only thing giving me pause on that argument today is the current RAM/NAND shortage
Not a shortage - price gouging. And it would mean an increase in the 'cloud' prices because they need to refresh the HW too. So by the summer the equation would be back to it.
> From a cache hierarchy standpoint, the design groups cores into four-core blocks that share approximately 4 MB of L2 cache per block. As a result, the aggregate last-level cache across the full package surpasses 1 GB, roughly 1,152 MB in total.
If cores are grouped into four-core blocks, and each block has 4MB of cache… isn’t that just 1MB per core? So 288MB total?
> these processors pack in up to 288 of the little guys as well as 576MB of last-level cache, 96 PCIe 5.0 lanes, and 12-channel DDR5-8000.
> The Xeon 6+ processors each have up to 12 compute tiles fabbed on 18A, all of which have six quad-core modules for a total of 24 cores per tile. There are also three 'active' base tiles on Intel 3, so-called because the base tiles include 192MB of last-level cache, which is so-called because each compute tile has 48MB of L3 cache.
So maybe 1MB per core L2, then 192MB of basically-L4 per base tile, then 48MB of L3 per compute tile? 192*3+48*12 gets me to the 1152, maybe that’s it.
Anyway, apparently these things will have “AMX” matrix extensions. I wonder if they’ll be good number crunchers.
With packages like this (lots of cores, multi-chip packaging, lots of memory channels), the architecture is increasingly a small cluster on a package rather than a monolithic CPU.
I wonder whether the next bottleneck becomes software scheduling rather than silicon - OS/runtimes weren’t really designed with hundreds of cores and complex interconnect topologies in mind.
Yes there are scheduling issues, Numa problems , etc caused by the cluster in a box form factor.
We had a massive performance issue a few years ago that we fixed by mapping our processes to the numa zones topology . The default design of our software would otherwise effectively route all memory accesses to the same numa zone and performance went down the drain.
There definitely are bottlenecks. The one I always think of is the kernel's networking stack. There's no sense in using the kernel TCP stack when you have hundreds of independent workloads. That doesn't make any more sense than it would have made 20 years ago to have an external TCP appliance at the top of your rack. Userspace protocol stacks win.
No they don't. They are horribly wasteful and inefficient compared to kernel TCP. Also they are useless because they sit on top of a kernel network interface anyways.
Unless you're doing specific tricks to minimize latency (HFT, I guess?) then there is no point.
Do the partitioned stacks of network namespaces share a single underlying global stack or are they fully independent instances? (And if not, could they be made so?)
I don't think there are any fundamental bottlenecks here. There's more scheduling overhead when you have a hundred processes on a single core than if you have a hundred processes on one hundred cores.
The bottlenecks are pretty much hardware-related - thermal, power, memory and other I/O. Because of this, you presumably never get true "288 core" performance out of this - as in, it's not going to mine Bitcoin 288 as fast as a single core. Instead, you have less context-switching overhead with 288 tasks that need to do stuff intermittently, which is how most hardware ends up being used anyway.
Maybe no fundamental bottlenecks but it's easy to accidentally write software that doesn't scale as linearly as it should, e.g. if there's suddenly more lock contention than you were expecting, or in a more extreme case if you have something that's O(n^2) in time or space, where n is core count.
afaik the mainline limit is 4096 threads. HP sells server with 32 sockets x 60 cores/socket x 2 threads/core = 3840 threads, so we are pretty close to that limit.
> OS/runtimes weren’t really designed with hundreds of cores and complex interconnect topologies in mind.
I mean....
IMO Erlang/Elixir is a not-terrible benchmark for how things should work in that state... Hell while not a runtime I'd argue Akka/Pekko on JVM Akka.Net on the .NET side would be able to do some good with it...[0] Similar for Go and channels (at least hypothetically...)
[0] - Of course, you can write good scaling code on JVM or CLR without these, but they at least give some decent guardrails for getting a good bit of the Erlang 'progress guaranteed' sauce.
> I wonder whether the next bottleneck becomes software scheduling rather than silicon
Yep, the scheduling has been a problem for a while. There was an amazing article few years ago about how the Linux kernel was accidentally hardcoded to 8 cores, you can probably google and find it.
IMO the most interesting problem right now is the cache, you get a cache miss every time a task is moving core. Problem, with thousands of threads switching between hundreds of cores every few milliseconds, we're dangerously approaching the point where all the time is spent trashing and reloading the CPU cache.
I think everyone's focusing on the core count, but the packaging story is way more interesting here. This thing is 12 separate chiplets on 18A stacked on base dies made on Intel 3, connected to I/O tiles on Intel 7. Three different process nodes in one package, shipping at volume. That's nuts.
And it's clearly an IFS play too. Intel Foundry needs a proof point — you can publish PDKs all day, but nothing sells foundry credibility like eating your own cooking in a 288-core server part at 450W. If Foveros Direct works here, it's the best ad Intel could run for potential foundry customers.
The chiplet sizing is smart for another reason nobody's mentioned: yield. 18A is brand new, yields are probably rough. But 24 cores per die is small enough that even bad yields give you enough good chiplets. Basically AMD's Zen playbook but with a 3D twist.
Also — 64 CXL 2.0 lanes! Several comments here are complaining about DDR5 prices, which is fair. But CXL memory pooling across a rack could change that math completely. I wonder if Intel is betting the real value isn't the cores but being the best CXL hub in the datacenter.
The ARM competition is still the elephant in the room though. "Many efficient cores" is what ARM has always done natively, and 17% IPC uplift on Darkmont doesn't close that gap by itself.
Agree entirely with your take. The packaging story is awesome, I wish there were more details on the stacking used on this one.
But I am at a loss to how Intel are really going to get any traction with IFS. How can anyone trust Intel as a long-term foundry partner. Even if they priced it more aggressively, the opportunity cost in picking a supplier who decides to quit next year would be catastrophic for many. The only way this works is if they practically give their services away to someone big, who can afford to take that risk and can also make it worth Intel's continued investment. Any ideas who that would be, I've got nothing.
I suspect that timing might help Intel here, with so much of the better established foundries near fully allocated for the next two years, it may be more a question of availability than brand name risk. And for whatever problems Intel has, it's pretty unlikely they'd go completely under and disolve in less than a year. Good non completion clauses in the contracts can mitigate a good chunk of the remaining risk.
Not to mention potential customers who would prefer a US based foundry regardless. My guess is that there's a pretty large part of the market that would be perfectly fine with using Intel.
Helped a friend make a difficult career decision (cozy job vs something hard and new + moving to a new city) that ultimately ended up with him working on the project. Glad that happened. I love to see people grow.
As a Yocto enthusiast, I am curious as to how much elapsed realtime would be needed for a clean Yocto build. Yocto is thread heavy, so with 288, it oughta be good.
My Yocto build times on a 32-core AMD are negligible, <2 minutes for a full distro, IIRC. I suspect higher core counts have diminishing returns, especially since most dev builds are heavily cached.
I’ve not kept up with Intel in a while, but one thing that stood out to me is these are all E cores— meaning no hyperthreading. Is something like this competitive, or preferred, in certain applications? Also does anyone know if there have been any benchmarks against AMDs 192 core Epyc CPU?
"Is something like this competitive, or preferred, in certain applications?"
They cite a very specific use case in the linked story: Virtualized RAN. This is using COTS hardware and software for the control plane for a 5G+ cell network operation. A large number of fast, low power cores would indeed suit such a application, where large numbers of network nodes are coordinated in near real time.
It's entirely possible that this is the key use case for this device: 5G networks are huge money makers and integrators will pay full retail for bulk quantities of such devices fresh out of the foundry.
In HPC, like physics simulation, they are preferred. There's almost no benefit of HT. What's also preferred is high cluck frequencies. These high core count CPUs nerd their clixk frequencies though.
It all depends on your exact workload, and I’ll wait to see benchmarks before making any confident claims, but in general if you have two threads of execution which are fine on an E-core, it’s better to actually put them on two E-cores than one hyperthreaded P-core.
Without the hyperthreading (E-cores) you get more consistent performance between running tasks, and cloud providers like this because they sell "vCPUs" that should not fluctuate when someone else starts a heavy workload.
Sort of. They can just sell even numbers of vCPUs, and dedicate each hyper-thread pair to the same tenant. That prevents another tenant from creating hyper-threading contention for you.
I don't know the nitty-gritty of why, but some compute intensive tasks don't benefit from hyperthreading. If the processor is destined for those tasks, you may as well use that silicon for something actually useful.
It's a few things; mostly along the lines of data caching (i.e. hyper threading may mean that other thread needs a cache sync/barrier/etc).
That said I'll point to the Intel Atom - the first version and refresh were an 'in-order' where hyper-threading was the cheapest option (both silicon and power-wise) to provide performance, however with Silvermont they switched to OOO execution but ditched hyper threading.
I think some of why is size on die. 288 E cores vs 72 P cores.
Also, there's so many hyperthreading vulnerabilities as of late they've disabled on hyperthreaded data center boards that I'd imagine this de-risks that entirely.
For an application like a build server, the only metric that really matters is total integer compute per dollar and per watt. When I compile e.g a Yocto project, I don't care whether a single core compiles a single C file in a millisecond or a minute; I care how fast the whole machine compiles what's probably hundred thousands of source files. If E-cores gives me more compute per dollar and watt than P-cores, give me E-cores.
Of course, having fewer faster cores does have the benefit that you require less RAM... Not a big deal before, you could get 512GB or 1TB of RAM fairly cheap, but these days it might actually matter? But then at the same time, if two E-cores are more powerful than one hyperthreaded P-core, maybe you actually save RAM by using E-cores? Hyperthreading is, after all, only a benefit if you spawn one compiler process per CPU thread rather than per core.
EDIT: Why in the world would someone downvote this perspective? I'm not even mad, just confused
I've seen scenarios where HT doesn't help, iirc very CPU-heavy things without much waiting on memory access. Which makes sense because the vcores are sharing the ALU.
Also have seen it disabled in academic settings where they want consistent performance when benchmarking stuff.
A bad moment to have a make-or-break moment for your CPU business - a lot of customers will probably hold off purchases right now because of the RAM prices, no matter how good your CPU might be.
Isn't this new server CPU a drop in replacement though? So the DC could pull off the old CPU, drop in the new one and not touch the existing RAM setup, yet be able to deliver better performance within the limits of the existing RAM. Then once RAM prices drop (okay that might be a while) separately upgrade the RAM at a different time.
That's semi-dependent on supplier arrangements; i.e. lots of shops won't want to upgrade CPUs on a server out of fear that they can't get support later; sometimes that's justified by contract, sometimes it's not.
Readit News
* Identify the workloads that haven't scaled in a year. Your ERPs, your HRIS, your dev/stage/test environments, DBs, Microsoft estate, core infrastructure, etc. (EDIT, from zbentley: also identify any cross-system processing where data will transfer from the cloud back to your private estate to be excluded, so you don't get murdered with egress charges)
* Run the cost analysis of reserved instances in AWS/Azure/GCP for those workloads over three years
* Do the same for one of these high-core "pizza boxes", but amortized over seven years
* Realize the savings to be had moving "fixed infra" back on-premises or into a colo versus sticking with a public cloud provider
Seriously, what took a full rack or two of 2U dual-socket servers just a decade ago can be replaced with three 2U boxes with full HA/clustering. It's insane.
Back in the late '10s, I made a case to my org at the time that a global hypervisor hardware refresh and accompanying VMware licenses would have an ROI of 2.5yrs versus comparable AWS infrastructure, even assuming a 50% YoY rate of license inflation (this was pre-Broadcom; nowadays, I'd be eyeballing Nutanix, Virtuozzo, Apache Cloudstack, or yes, even Proxmox, assuming we weren't already a Microsoft shop w/ Hyper-V) - and give us an additional 20% headroom to boot. The only thing giving me pause on that argument today is the current RAM/NAND shortage, but even that's (hopefully) temporary - and doesn't hurt the orgs who built around a longer timeline with the option for an additional support runway (like the three-year extended support contracts available through VARs).
If we can't bill a customer for it, and it's not scaling regularly, then it shouldn't be in the public cloud. That's my take, anyway. It sucks the wind from the sails of folks gung-ho on the "fringe benefits" of public cloud spend (box seats, junkets, conference tickets, etc...), but the finance teams tend to love such clear numbers.
For those that do, your scaling example works against you. If today you can merge three services into one, then why do you need full time infrastructure staff to manage so few servers? And remember, you want 24/7 monitoring, replication for disaster recovery, etc. Most businesses do not have IT infrastructure as a core skill or differentiator, and so they want to farm it out.
This is really the core problem. Every time I’ve done the math on a sizable cloud vs on-prem deployment, there is so much money left on the table that the orgs can afford to pay FAANG-level salaries for several good SREs but never have we been able to find people to fill the roles or even know if we had found them.
The numbers are so much worse now with GPUs. The cost of reserved instances (let alone on-demand) for an 8x H100 pod even with NVIDIA Enterprise licenses included leaves tens of thousands per pod for the salary of employees managing it. Assuming one SREs can manage at least four racks the hardware pays for itself, if you can find even a single qualified person.
The company did need the same exact people to manage AWS anyway. And the cost difference was so high that it was possible to hire 5 more people which wasn't needed anyway.
Not only the cost but not needing to worry about going over the bandwidth limit and having soo much extra compute power made a very big difference.
Imo the cloud stuff is just too full of itself if you are trying to solve a problem that requires compute like hosting databases or similar. Just renting a machine from a provider like Hetzner and starting from there is the best option by far.
That's partially true; managing cloud also takes skill, most people forget that with end result being "well we saved on hiring sysadmins, but had to have more devops guys". Hell I manage mostly physical infrastructure (few racks, few hundred VMs) and good 80% of my work is completely unrelated to that, it's just the devops gluing stuff together and helping developers to set their stuff up, which isn't all that different than it would be in cloud.
> And remember, you want 24/7 monitoring, replication for disaster recovery, etc.
And remember, you need that for cloud too. Plenty of cloud disaster stories to see where they copy pasted some tutorial thinking that's enough then surprise.
There is also partial way of just getting some dedicated servers from say OVH and run infra on that, you cut out a bit of the hardware management from skillset and you don't have the CAPEX to deal with.
But yes, if it is less than at least a rack, it's probably not worth looking for onprem unless you have really specific use case that is much cheaper there (I mean less than usual half)
On top of that no one really knows what the fuck they are doing in AWS anyway.
Not quite. If you hire a bad talent to manage your 'cloud gear' then you would find what the mistakes which would cost you nothing on-premises would cost you in the cloud. Sometimes - a lot.
Is it still a problem in 2026 when unemployment in IT is rising? Reasons can be argued (the end of ZIRP or AI) but hiring should be easier than it was at any time during the last 10 years.
This comes up again and again. It was the original sales pitch from cloud vendors.
Often the very same companies repeating this messaging are recruiting and paying large teams of platform developers to manage their cloud…and pay for them to be on call.
Generalists (it me) typically command lower market rates, remain far more flexible, and can replicate much of the experience on-prem while knowing when and why to put something in the public cloud. A few examples:
* 24/7 Telemetry and Monitoring: if you have the talent to roll your own with OpenTelemetry, Grafana, Prometheus, and a database to store the telemetry involved, then great! If it’s me wrangling a hybrid environment though, I’m likely leaning on New Relic to save on headcount and deliver similar results.
* DBaaS: this is increasingly just offered by hypervisor managers since it’s often just spooling up a container and pointing to storage. Not quite a “solved problem”, but enough of one that a single DBA can cover both estates if needed - or a moderately skilled generalist can at least secure it for internal use before offering it out to customers
* Vulnerability Management: as much as I’d love to have an internal Red Team, it’s an order of magnitude cheaper to leverage Nessus, Wazuh, Wiz, or any of the other fleets of continuous scanners to identify vulnerabilities or misconfigurations
That’s just me cherry-picking. The point is less “do everything possible on-prem”, and more “diversify workload placement depending on cost advantages relative to risk models”, and on-prem wins quite handily for a lot of LOB software that just quietly sits and does its thing with minimal fuss.
I’ve had success with this approach by keeping it to only the business process management stacks (CRMs, AD, and so on—examples just like the ones you listed). But as soon as there’s any need for bridging cloud/onprem for any data rate beyond “cronned sync” or “metadata only”, it starts to hurt a lot sooner than you’d expect, I’ve found.
Folks wanting one or the other miss savings had by effectively leveraging both.
Its hard drive and SSD space prices that stagger me on the cloud. Where one of the server CPUs might only be about 2x the price of buy a CPU for a few years if you buy less in a small system (all be it with less clock speed usually on the cloud) the drive space is at least 10-100x the price of doing it locally. Its got a bit more potential redudency but for that overhead you can repeat that data a lot of times.
As time has gone on the deal of cloud has got worse as the hardware got more cores.
That makes a lot of sense. Cloud providers are selling compute, and as cores get faster, the single core gets more expensive.
If that's you then the GraniteRapids AP platform that launched previously to this can hit similar numbers of threads (256 for the 6980P). There are a couple of caveats to this though - firstly that there are "only" 128 physical cores and if you're using VMs you probably don't want to share a physical core across VMs, secondly that it has a 500W TDP and retails north of $17000, if you can even find one for sale.
Overall once you're really comparing like to like, especially when you start trying to have 100+GbE networking and so on, it gets a lot harder to beat cloud providers - yes they have a nice fat markup but they're also paying a lot less for the hardware than you will be.
Most of the time when I see takes like this it's because the org has all these fast, modern CPUs for applications that get barely any real load, and the machines are mostly sitting idle on networks that can never handle 1/100th of the traffic the machine is capable of delivering. Solving that is largely a non-technical problem not a "cloud is bad" problem.
Darkmont is a slightly improved variant of the Skymont cores used in Arrow Lake/Lunar Lake and it has a performance very similar to the Arm Neoverse V3 cores used in Graviton5, the latest generation of custom AWS CPUs.
However, a Clearwater Forest Xeon CPU has much more cores per socket than Graviton5 and it also supports dual-socket motherboards.
Darkmont also has a greater performance than the older big Intel cores, like all Skylake derivatives, inclusive for AVX-using programs, so it is no longer comparable with the Atom series of cores from which it has evolved.
Darkmont is not competitive in absolute performance with AMD Zen 5, but for the programs that do not use AVX-512 it has better performance per watt.
However, since AMD has started to offer AVX-512 for the masses, the number of programs that have been updated to be able to benefit from AVX-512 is increasing steadily, and among them are also applications where it was not obvious that using array operations may enhance performance.
Because of this pressure from AMD, it seems that this Clearwater Forest Xeon is the final product from Intel that does not support AVX-512. Both next 2 Intel CPUs support AVX-512, i.e. the Diamond Rapids Xeon, which might be launched before the end of the year, and the desktop and laptop CPU Nova Lake, whose launch has been delayed to next year (together with the desktop Zen 6, presumably due to the shortage of memories and production allocations at TSMC).
At my job we use HyperV, and finding someone who actually knows HyperV is difficult and expensive. Throw in Cisco networking, storage appliances, etc to make it 99.99% uptime...
Also that means you have just one person, you need at least two if you don't want gaps in staffing, more likely three.
Then you still need all the cloud folks to run that.
We have a hybrid setup like this, and you do get a bit of best of both worlds, but ultimately managing onprem or colo infra is a huge pain in the ass. We only do it due to our business environment.
All of the complexity of onprem, especially when you need to worry about failover/etc can get tricky, especially if you are in a wintel env like a lot of shops are.
i.e. lots of companies are doing sloppy 'just move the box to an EC2 instance' migrations because of how VMWare jacked their pricing up, and now suddenly EC2/EBS/etc costing is so cheap it's a no brain choice.
I think the knowledge base to set up a minimal cost solution is too tricky to find a benefit vs all the layers (as you almost touched on, all the licensing at every layer vs a cloud provider managing...)
That said, rug pulls are still a risk; I try to push for 'agnostic' workloads in architecture, if nothing else because I've seen too many cases where SaaS/PaaS/etc decide to jack up the price of a service that was cheap, and sure you could have done your own thing agnostically, but now you're there, and migrating away has a new cost.
IOW, I agree; I don't think the human capital is there as far as infra folks who know how to properly set up such environments, especially hitting the 'secure+productive' side of the triangle.
> At my job we use HyperV, and finding someone who actually knows HyperV is difficult and expensive...
Try offering significantly higher pay.
AMD has had these sorts of densities available for a minute.
> Identify the workloads that haven't scaled in a year.
I have done this math recently, and you need to stop cherry picking and move everything. And build a redundant data center to boot.
Compute is NOT the major issue for this sort of move:
Switching and bandwidth will be major costs. 400gb is a minimum for interconnects and for most orgs you are going to need at least that much bandwidth top of rack.
Storage remains problematic. You might be able to amortize compute over this time scale, but not storage. 5 years would be pushing it (depending on use). And data center storage at scale was expensive before the recent price spike. Spinning rust is viable for some tasks (backup) but will not cut it for others.
Human capital: Figuring out how to support the hardware you own is going to be far more expensive than you think. You need to expect failures and staff accordingly, that means resources who are going to be, for the most part, idle.
Im pretty sure a box like this could run our whole startup, hosting PG, k8s, our backend apis, etc, would be way easier to setup, and not cost 2 devops and $40,000 a month to do it.
- request some HW to run $service
- the "IT dept" (really, self-interested gatekeeper) might give you something now, or in two weeks, or god help you if they need to order new hardware then its in two months, best case
- there will be various weird rules on how the on-prem HW is run, who has access etc, hindering developer productivity even further
- the hardware might get insanely oversubscribed so your service gets half a cpu core with 1GB RAM, because perverse incentives mean the "IT dept" gets rewarded for minimizing cost, while the price is paid by someone else
- and so on...
The cloud is a way around this political minefield.
Until the bills _really_ start skyrocketing...
>Anyway, I wasn't able to test using proxmox or vmware on there to split up cpu/memory resources; we decided instead to just buy a bunch of smaller-core-count AMD Ryzen 1Us instead, which scaled way better with my naive approac
If that was single 384 (192 times 2 for hyperthreading) CPU you are getting "only" 12 DDR5 channels, so one RAM channel is shared by 16c/32y
So just plain 16 core desktop Ryzen will have double memory bandwidth per core
And 384 actual cores or 384 hyperthreading cores?
Inference is so memory bandwidth heavy that my expectations are low. An EPYC getting 12 memory channels instead of 2 only goes so far when it has 24x as many cores.
The core density is bullshit when each core is so slow that it can't do any meaningful work. The reality is that Intel is 3 times behind AMD/TSMC on performance vs power consumption ratio.
People would be better off having a look at the high frequency models (9xx5F models like the 9575F), that was the first generation of CPU server to reach ~5 GHz and sustain it on 32+ cores.
I agree, but.
For one, it's not just the machines themselves. You also need to budget in power, cooling, space, the cost of providing redundant connectivity and side gear (e.g. routers, firewalls, UPS).
Then, you need a second site, no matter what. At least for backups, ideally as a full failover. Either your second site is some sort of cloud, which can be a PITA to set up without introducing security risks, or a second physical site, which means double the expenses.
If you're a publicly listed company, or live in jurisdictions like Europe, or you want to have cybersecurity insurance, you have data retention, GDPR, SOX and a whole bunch of other compliance to worry about as well. Sure, you can do that on-prem, but you'll have a much harder time explaining to auditors how your system works when it's a bunch of on-prem stuff vs. "here's our AWS Backup plans covering all servers and other data sources, here is the immutability stuff, here are plans how we prevent backup expiry aka legal hold".
Then, all of that needs to be maintained, which means additional staff on payroll, if you own the stuff outright your finance team will whine about depreciation and capex, and you need to have vendors on support contracts just to get firmware updates and timely exchanges for hardware under warranty.
Long story short, as much as I prefer on-prem hardware vs the cloud, particularly given current political tensions - unless you are a 200+ employee shop, the overhead associated with on-prem infrastructure isn't worth it.
You can technically have backblaze's unlimited backup option which costs around 7$ for a given machine although its more intended for windows, there have been people who make it work and Daily backups and it should work with gdpr (https://www.backblaze.com/company/policy/gdpr) with something like hetzner perhaps if you are worried about gdpr too much and OVH storage boxes (36 TB iirc for ~55$ is a good backup box) and you should try to follow 3-2-1 strategy.
> Then, all of that needs to be maintained, which means additional staff on payroll, if you own the stuff outright your finance team will whine about depreciation and capex, and you need to have vendors on support contracts just to get firmware updates and timely exchanges for hardware under warranty.
I can't speak for certain but its absolutely possible to have something but iirc for companies like dell, its possible to have products be available on a monthly basis available too and you can simply colocate into a decent datacenter. Plus points in that now you can get 10-50 GB ports as well if you are too bandwidth hungry and are available for a lot lot more customizable and the hardware is already pretty nice as GP observed. (Yes Ram prices are high, lets hope that is temporary as GP noted too)
I can't speak about firmware updates or timely exchanges for hardware under security.
That being said, I am not saying this is for everyone as well. It does essentially boils down to if they have expertise in this field/can get expertise in this field or not for cheaper than their aws bills or not. With many large AWS bills being in 10's of thousands of dollars if not hundreds of thousands of dollars, I think that far more companies might be better off with the above strategy than AWS actually.
Deleted Comment
As other people have pointed out: what happens when the PSU or mobo shits itself, what happens when the new version of vmware or docker (or whatever) shits itself, etc
cloud = opex
The accounting dept will always win this debate.
Context: been helpdesk, sysadmin & network admin, DevOps, Site Reliability Engineer, in that progression, starting in the 90's. Max on-prem was 40 racks, scaled up and down over years.
Many comments talking about how staffing is a key element of this equation that can't be overlooked, but I decided to reply to the root comment, which doesn't say whether/how it considers staffing.
This is a complex equation - and it is relatively easy to present an incomplete or misleading picture management to push the move into the cloud.. or out of the cloud.
Some factors, in no particular order:
1) Scaling: it is self-evident that pulling a single physical server worth out of the cloud is not worth it.. even for 288 cores. Or perhaps 1152 for 4xXeon in a single server. Still likely not worth it. Why? Because a single server is never just that. Someone has to swap components when it goes down. When it goes down.. ALL 1152 cores are down, along with everything they are doing. Is that acceptable for all applications running on all those cores? It is also appropriate supporting infrastructure - power, cooling, physical space. The "fairly obvious" minimum scaling is "enough servers that one can be entirely down for maintenance while keeping everything else running." But now you're paying for some overhead. At 2 servers, you're buying 2x what you need, half that capacity is idle all the time. And so on.
On this point - I think the other comments talking about "each SRE managing 4 (or 5, or 7)" racks missed the point entirely. SRE's should be doing scalable work, whether in the cloud, or on-prem. And they should NOT be swapping failed hard drives and power supplies. Designing a larger-than-one rack install is probably worth hiring consultants for if you don't have that expertise in-house, though the SREs that would be supporting it would need to supply lots of input. To some extent, server & network equipment vendors can also help. It is not trivial as the scale goes up. But then it should run for some years, with relatively unskilled people handling hardware failures and you can re-engage consultants if necessary to do upgrades as hardware and needs evolve.
But your SREs should be on-staff, and probably on-call to handle the software running on that hardware.. and to some extent to call the remote hands to deal with hardware failures.
2) Business needs: does the business need the tech skills that self-hosting requires for the core business? For example - if the business itself is cloud SAAS, maybe DIY-ing at least some of your infrastructure is right in your wheelhouse. If so - a modest increase in staff could mean a huge cost savings. But if not, all the cost of skilled staff to run it is simply part of the cost of in-housing this stuff.
3) Staffing: the people that swap broken hardware are not the same people that respond to pages because the business critical application crashed due to a bug. You can pay a colo facility for all this, typically by the hour - but it isn't cheap and you've got to supply all the spares etc. Is that part of your budget for on-prem?
4) On-call: maybe your self-hosted ERP system can be down every night and weekend without issues.. and even business hours can tolerate 98% uptime. But that doesn't mean you can get away without having someone on call - presumably you're hosting more than just this lowish-requirement ERP system. I'll disagree with other comments - the "no burnout" number of on-call staff you need is 6-7, not 4! Remember people take vacations too. This is well studied, and established, I'll reference Tom Limoncelli's books. This could be relatively cheap and require fewer staff with geo-distributed staff, and it would tend to overlap with staff you already use to provide on-call for anything you host on the cloud - so maybe for your situation it is close to a wash. But you can't forget to budget for it even if the line item is $0.
5) Vendor support: maybe you already have your own data center or colo and are hosting a ton of stuff. Why not move all your Atlassian stuff in house and save the hosting cost. Oh.. wups, Atlassian simply doesn't support that any more. Host it with Atlassian or GTFO. A minor point as most vendors would give you enough notice you can simply run out the lifetime of the hardware it is on and not replace it.
6) Market pricing: At one point Amazon was starting "by the minute, by the core cloud" (as opposed the older "cloudish" model of leasing only an entire physical server by the entire year) and priced a bit under market to get going. Then once they established dominance, they cranked the price WAY up for profit extraction. But now they do have some competition and they're a bit more selective about how they extract profits. In my perception they've shifted a lot of the profit taking to the value-added services rather than raw instance time, but I could be wrong. And they have HUGE costs - it is beyond naive to look at the per-hour cost of an instance and compare it to purchasing an identical physical server solely on the purchase price of that server. Corey Quinn / Duckbill group has spent a huge part of his career in this space - if you're already in the cloud 100% it is well worth optimizing those costs before you start comparing it to what on-prem might cost.
Not a shortage - price gouging. And it would mean an increase in the 'cloud' prices because they need to refresh the HW too. So by the summer the equation would be back to it.
> From a cache hierarchy standpoint, the design groups cores into four-core blocks that share approximately 4 MB of L2 cache per block. As a result, the aggregate last-level cache across the full package surpasses 1 GB, roughly 1,152 MB in total.
If cores are grouped into four-core blocks, and each block has 4MB of cache… isn’t that just 1MB per core? So 288MB total?
HotHardware reports
https://hothardware.com/news/intel-clearwater-forest-xeon-6-...
> these processors pack in up to 288 of the little guys as well as 576MB of last-level cache, 96 PCIe 5.0 lanes, and 12-channel DDR5-8000.
> The Xeon 6+ processors each have up to 12 compute tiles fabbed on 18A, all of which have six quad-core modules for a total of 24 cores per tile. There are also three 'active' base tiles on Intel 3, so-called because the base tiles include 192MB of last-level cache, which is so-called because each compute tile has 48MB of L3 cache.
So maybe 1MB per core L2, then 192MB of basically-L4 per base tile, then 48MB of L3 per compute tile? 192*3+48*12 gets me to the 1152, maybe that’s it.
Anyway, apparently these things will have “AMX” matrix extensions. I wonder if they’ll be good number crunchers.
I wonder whether the next bottleneck becomes software scheduling rather than silicon - OS/runtimes weren’t really designed with hundreds of cores and complex interconnect topologies in mind.
We had a massive performance issue a few years ago that we fixed by mapping our processes to the numa zones topology . The default design of our software would otherwise effectively route all memory accesses to the same numa zone and performance went down the drain.
No they don't. They are horribly wasteful and inefficient compared to kernel TCP. Also they are useless because they sit on top of a kernel network interface anyways.
Unless you're doing specific tricks to minimize latency (HFT, I guess?) then there is no point.
The bottlenecks are pretty much hardware-related - thermal, power, memory and other I/O. Because of this, you presumably never get true "288 core" performance out of this - as in, it's not going to mine Bitcoin 288 as fast as a single core. Instead, you have less context-switching overhead with 288 tasks that need to do stuff intermittently, which is how most hardware ends up being used anyway.
You memory only has so much bandwidth, but now it's shared by even more cores.
So on the OS side we might already have the needed tools for these CoC (cluster on chip ;))
But that's just one piece of the puzzle, I guess.
I mean....
IMO Erlang/Elixir is a not-terrible benchmark for how things should work in that state... Hell while not a runtime I'd argue Akka/Pekko on JVM Akka.Net on the .NET side would be able to do some good with it...[0] Similar for Go and channels (at least hypothetically...)
[0] - Of course, you can write good scaling code on JVM or CLR without these, but they at least give some decent guardrails for getting a good bit of the Erlang 'progress guaranteed' sauce.
Yep, the scheduling has been a problem for a while. There was an amazing article few years ago about how the Linux kernel was accidentally hardcoded to 8 cores, you can probably google and find it.
IMO the most interesting problem right now is the cache, you get a cache miss every time a task is moving core. Problem, with thousands of threads switching between hundreds of cores every few milliseconds, we're dangerously approaching the point where all the time is spent trashing and reloading the CPU cache.
https://news.ycombinator.com/item?id=38260935
> This article is clickbait and in no way has the kernel been hardcoded to a maximum of 8 cores.
And it's clearly an IFS play too. Intel Foundry needs a proof point — you can publish PDKs all day, but nothing sells foundry credibility like eating your own cooking in a 288-core server part at 450W. If Foveros Direct works here, it's the best ad Intel could run for potential foundry customers.
The chiplet sizing is smart for another reason nobody's mentioned: yield. 18A is brand new, yields are probably rough. But 24 cores per die is small enough that even bad yields give you enough good chiplets. Basically AMD's Zen playbook but with a 3D twist.
Also — 64 CXL 2.0 lanes! Several comments here are complaining about DDR5 prices, which is fair. But CXL memory pooling across a rack could change that math completely. I wonder if Intel is betting the real value isn't the cores but being the best CXL hub in the datacenter.
The ARM competition is still the elephant in the room though. "Many efficient cores" is what ARM has always done natively, and 17% IPC uplift on Darkmont doesn't close that gap by itself.
But I am at a loss to how Intel are really going to get any traction with IFS. How can anyone trust Intel as a long-term foundry partner. Even if they priced it more aggressively, the opportunity cost in picking a supplier who decides to quit next year would be catastrophic for many. The only way this works is if they practically give their services away to someone big, who can afford to take that risk and can also make it worth Intel's continued investment. Any ideas who that would be, I've got nothing.
Not to mention potential customers who would prefer a US based foundry regardless. My guess is that there's a pretty large part of the market that would be perfectly fine with using Intel.
With the standard form of business trust: a contract.
That the CPU cores are low frequency cores probably helps with yield as well.
They cite a very specific use case in the linked story: Virtualized RAN. This is using COTS hardware and software for the control plane for a 5G+ cell network operation. A large number of fast, low power cores would indeed suit such a application, where large numbers of network nodes are coordinated in near real time.
It's entirely possible that this is the key use case for this device: 5G networks are huge money makers and integrators will pay full retail for bulk quantities of such devices fresh out of the foundry.
https://www.comsol.com/support/knowledgebase/1096
That said I'll point to the Intel Atom - the first version and refresh were an 'in-order' where hyper-threading was the cheapest option (both silicon and power-wise) to provide performance, however with Silvermont they switched to OOO execution but ditched hyper threading.
Gaming CPUs and some EPYCs are the best
Also, there's so many hyperthreading vulnerabilities as of late they've disabled on hyperthreaded data center boards that I'd imagine this de-risks that entirely.
Of course, having fewer faster cores does have the benefit that you require less RAM... Not a big deal before, you could get 512GB or 1TB of RAM fairly cheap, but these days it might actually matter? But then at the same time, if two E-cores are more powerful than one hyperthreaded P-core, maybe you actually save RAM by using E-cores? Hyperthreading is, after all, only a benefit if you spawn one compiler process per CPU thread rather than per core.
EDIT: Why in the world would someone downvote this perspective? I'm not even mad, just confused
I imagine that means less C++/Rust than most, which means much less time spent serialized on the linker / cross compilation unit optimizer.
Which is why I used AMD in my last desktop computer build
Also have seen it disabled in academic settings where they want consistent performance when benchmarking stuff.
As to E core itself - it's ARM's playbook.