The Intel 12th Gen Core i9-12900K Review: Hybrid Performance Brings Hybrid Complexity
by Dr. Ian Cutress & Andrei Frumusanu on November 4, 2021 9:00 AM ESTCPU Tests: SPEC ST Performance on P-Cores & E-Cores
SPEC2017 is a series of standardized tests used to probe the overall performance between different systems, different architectures, different microarchitectures, and setups. The code has to be compiled, and then the results can be submitted to an online database for comparison. It covers a range of integer and floating point workloads, and can be very optimized for each CPU, so it is important to check how the benchmarks are being compiled and run.
For compilers, we use LLVM both for C/C++ and Fortan tests, and for Fortran we’re using the Flang compiler. The rationale of using LLVM over GCC is better cross-platform comparisons to platforms that have only have LLVM support and future articles where we’ll investigate this aspect more. We’re not considering closed-sourced compilers such as MSVC or ICC.
clang version 10.0.0
clang version 7.0.1 (ssh://git@github.com/flang-compiler/flang-driver.git
24bd54da5c41af04838bbe7b68f830840d47fc03)
-Ofast -fomit-frame-pointer
-march=x86-64
-mtune=core-avx2
-mfma -mavx -mavx2
Our compiler flags are straightforward, with basic –Ofast and relevant ISA switches to allow for AVX2 instructions. We decided to build our SPEC binaries on AVX2, which puts a limit on Haswell as how old we can go before the testing will fall over. This also means we don’t have AVX512 binaries, primarily because in order to get the best performance, the AVX-512 intrinsic should be packed by a proper expert, as with our AVX-512 benchmark. All of the major vendors, AMD, Intel, and Arm, all support the way in which we are testing SPEC.
To note, the requirements for the SPEC licence state that any benchmark results from SPEC have to be labeled ‘estimated’ until they are verified on the SPEC website as a meaningful representation of the expected performance. This is most often done by the big companies and OEMs to showcase performance to customers, however is quite over the top for what we do as reviewers.
For Alder Lake, we start off with a comparison of the Golden Cove cores, both in DDR5 as well as DDR4 variants. We’re pitting them as direct comparison against Rocket Lake’s Cypress Cove cores, as well as AMD’s Zen3.
Starting off in SPECint2017, the first thing I’d say is that for single-thread workloads, it seems that DDR5 doesn’t showcase any major improvements over DDR4. The biggest increase for the Golden Cove cores are in 520.omnetpp_r at 9.2% - the workload is defined by sparse memory accessing in a parallel way, so DDR5’s doubled up channel count here is likely what’s affecting the test the most.
Comparing the DDR5 results against RKL’s WLC cores, ADL’s GLC showcases some large advantages in several workloads: 24% in perlbench, +29% in omnetpp, +21% in xalancbmk, and +26% in exchange2 – all of the workloads here are likely boosted by the new core’s larger out of order window which has grown to up to 512 instructions. Perlbench is more heavily instruction pressure biased, at least compared to other workloads in the suite, so the new 6-wide decoder also likely is a big reason we see such a large increase.
The smallest increases are in mcf, which is more pure memory latency bound, and deepsjeng and leela, the latter which is particularly branch mispredict heavy. Whilst Golden Cove improves its branch predictors, the core also had to add an additional cycle of misprediction penalty, so the relative smaller increases here make sense with that as a context.
In the FP suite, the DDR5 results have a few larger outliers compared to the DDR4 set, bwaves and fotonik3d showcase +15% and +17% just due to the memory change, which is no surprise given both workloads extremely heavy memory bandwidth characteristic.
Compared to RKL, ADL showcases also some very large gains in some of the workloads, +33% in cactuBBSN, +24% in povray. The latter is a surprise to me as it should be a more execution-bound workload, so maybe the new added FADD units of the cores are coming into play here.
We’ve had not too much time to test out the Gracemont cores in isolation, but we are able to showcase some results. This set here is done on native Linux rather than WSL due to affinity issues on Windows, the results are within margin of error between the platforms, however there are a few % points outliers on the FP suite. Still, comparing the P to E-cores are in apples-to-apples conditions in these set of graphs:
When Intel mentioned that the Gracemont E-cores of Alder Lake were matching the ST performance of the original Skylake, Intel was very much correct in that description. Unlike what we consider “little” cores in a normal big.LITTLE setup, the E-cores of Alder Lake are still quite performant.
In the aggregate scores, an E-core is roughly 54-64% of a P-core, however this percentage can go as high as 65-73%. Given the die size differences between the two microarchitectures, and the fact that in multi-threaded scenarios the P-cores would normally have to clock down anyway because of power limits, it’s pretty evident how Intel’s setup with efficiency and density cores allows for much higher performance within a given die size and power envelope.
In SPEC, in terms of package power, the P-cores averaged 25.3W in the integer suite and 29.2W in the FP suite, in contrast to respectively 10.7W and 11.5W for the E-cores, both under single-threaded scenarios. Idle package power ran in at 1.9W.
Alder Lake and the Golden Cove cores are able to reclaim the single-threaded performance crown from AMD and Apple. The increases over Rocket Lake come in at +18-20%, and Intel’s advantage over AMD is now at 6.4% and 16.1% depending on the suite, maybe closer than what Intel would have liked given V-cache variants of Zen3 are just a few months away.
Again, the E-core performance of ADL is impressive, while not extraordinary ahead in the FP suite, they can match the performance of some middle-stack Zen2 CPUs from only a couple of years ago in the integer suite.
Facebook
Twitter
RSS
474 Comments
View All Comments
Oxford Guy - Sunday, November 7, 2021 - link
‘or maybe switch off their E-cores and enable AVX-512 in BIOS’This from exactly the same person who posted, just a few hours ago, that it’s correct to note that that option can disappear and/or be rendered non-functional.
I am reminded of your contradictory posts about ECC where you mocked advocacy for it (‘advocacy’ being merely its mention) and proceeded to claim you ‘wish’ for more ECC support.
Once again, it’s helpful to have a grasp of what one actually believes prior to posting. Allocating less effort to posting puerile insults and more toward substance is advised.
mode_13h - Sunday, November 7, 2021 - link
> This from exactly the same person who posted, just a few hours ago, that it’s> correct to note that that option can disappear and/or be rendered non-functional.
You need to learn to distinguish between what Intel has actually stated vs. the facts as we wish them to be. In the previous post you reference, I affirmed your acknowledgement that the capability disappearing would be consistent with what Intel has actually said, to date.
In the post above, I was leaving open the possibility that *maybe* Intel is actually "cool" with there being a BIOS option to trade AVX-512 for E-cores. We simply don't know how Intel feels about that, because (to my knowledge) they haven't said.
When I clarify the facts as they stand, don't confuse that with my position on the facts as I wish them to be. I can simultaneously acknowledge one reality, which maintaining my own personal preference for a different reality.
This is exactly what happened with the ECC situation: I was clarifying Intel's practice, because your post indicated uncertainty about that fact. It was not meant to convey my personal preference, which I later added with a follow-on post.
Having to clarify this to an "Oxford Guy" seems a bit surprising, unless you meant like Oxford Mississippi.
> you mocked advocacy
It wasn't mocking. It was clarification. And your post seemed more to express befuddlement than expressive of advocacy. It's now clear that your post was a poorly-executed attempt at sarcasm.
Once again, it's helpful not to have your ego so wrapped up in your posts that you overreact when someone tries to offer a factual clarification.
Oxford Guy - Monday, November 8, 2021 - link
I now skip to the bottom of your posts If I see more of the same preening and posing, I spare myself the rest of the nonsense.mode_13h - Tuesday, November 9, 2021 - link
> If I see more of the same preening and posing, I spare myself the rest of the nonsense.Then I suggest you don't read your own posts.
I can see that you're highly resistant to reason and logic. Whenever I make a reasoned reply, you always hit back with some kind of vague meta-critique. If that's all you've got, it can be seen as nothing less than a concession.
O-o-o-O - Saturday, November 6, 2021 - link
Anyone talking about dumping x64 ISA?I don't see AVX-512 a good solution. Current x64 chips are putting so much complexity in CPU with irrational clock speed that migrating process-node further into Intel4 on would be a nightmare once again.
I believe most of the companies with in-house developers expect the end of Xeon-era is quite near, as most of the heavy computational tasks are fully optimized for GPUs and that you don't want coal burning CPUs.
Even if it doesn't come in 5 year time-frame, there's a real threat and have to be ahead of time. After all, x86 already extended its life 10+ years when it could have been discontinued. Now it's really a dinosaur. If so, non-server applications would follow the route as well.
We want more simple / solid / robust base with scalability. Not an unreliable boost button that sometimes do the trick.
SystemsBuilder - Saturday, November 6, 2021 - link
I don't see AVX-512 that negatively it is just the same as AVX2 but double the vectors size and a with a richer instruction set. I find it pretty cool to work with especially when you've written some libraries that can take advantage of it. As I wrote before, it looks like Golden cove got AVX-512 right based on what Ian and Andrei uncovered. 0 negative offset (e.g. running at full speed), power consumption not much more than AVX2, and it supports both FP16 and BP16 vectors! I think that's pretty darn good! I can work with that! Now I want my Sapphire rapids with 32 or 48 Golden cove P cores! No not fall 2022 i want it now! lolmode_13h - Saturday, November 6, 2021 - link
> When you optimize code today (for pre Alder lake CPUs) to take advantage> of AVX-512 you need to write two paths (at least).
Ah, so your solution depends on application software changes, specifically requiring them to do more work. That's not viable for the timeframe of concern. And especially not if its successor is just going to add AVX-512 to the E-cores, within a year or so.
> There are many levels of AVX-512 support and effectively you need write customized
> code for each specific CPUID
But you don't expect the capabilities to change as a function of which thread is running, or within a program's lifetime! What you're proposing is very different. You're proposing to change the ABI. That's a big deal!
> It is absolutely possible and it will come with time.
Or not. ARM's SVE is a much better solution.
> I think in the future P and E cores might have more than just AVX-512 that is different
On Linux, using AMX will require a thread to "enable" it. This is a little like what you're talking about. AMX is a big feature, though, and unlike anything else. I don't expect to start having to enable every new ISA extension I want to use, or query how many hyperthreads actually support - this becomes a mess when you start dealing with different libraries that have these requirements and limitations.
Intel's solution isn't great, but it's understandable and it works. And, in spite of it, they still delivered a really nice-performing CPU. I think it's great if technically astute users have/retain the option to trade E-cores for AVX-512 (via BIOS), but I think it's kicking a hornets nest to go down the path of having a CPU with asymmetrical capabilities among its cores.
Hopefully, Raptor Lake just adds AVX-512 to the E-cores and we can just let this issue fade into the mists of time, like other missteps Intel & others have made.
SystemsBuilder - Saturday, November 6, 2021 - link
I too believe AVX-512 exclusion in the E cores it is transitory. next gen E cores may include it and the issue goes away for AVX-512 at least (Raptor Lake?). Still there will be other features that P have but E won't have so the scheduler needs to be adjusted for that. This will continue to evolve with every generation of E and P cores - because they are here to stay.I read somewhere a few months ago but right now i do not remember where (maybe on Anandtech not sure) that the AVX-512 transistor budget is quite small (someone measured it on the die) so not really a big issue in terms of area.
AMX is interesting because where AVX-512 are 512 bit vectors, AMX is making that 512x512 bit matrices or tiles as intel calls it. Reading the spec on AMX you have BF16 tiles which is awesome if you're into neural nets. Of course gpus will still perform better with matrix calculations (multiplications) but the benefit with AMX is that you can keep both the general CPU code and the matrix specific code inside the CPU and can mix the code seamlessly and that's gonna be very cool - you cut out the latency between GPU and CPU (and no special GPU API's are needed). but of course you can still use the GPU when needed (sometimes it maybe faster to just do a matrix- matrix add for instance just inside the CPU with the AMX tiles) - more flexibility.
Anyway, I do think we will run into a similar issue with AMX as we have the AVX-512 on Alder Lake and therefore again the scheduler needs to become aware of each cores capabilities and each piece of code need to state what type of core they prefer to run on: AVX2, AVX-512, AMX capable core etc (the compliers job). This way the scheduler can do the best job possible with every thread.
There will be some teething for a while but i think this is the direction it is going.
mode_13h - Sunday, November 7, 2021 - link
The difference is that AMX is new. It's also much more specialized, as you point out. But that means that they can place new hoops for code to jump through, in order to use it.It's very hard to put a cat like AVX-512 back in the bag.
SystemsBuilder - Saturday, November 6, 2021 - link
To be clear, I also want to add that the way code is written today (in my organization) pre Alder Lake code base. Every time we write a code path for AVX512 we need to write a fallback code path incase the CPU is not AVX-512 capable. This is standard (unless you can control the execution H/W 100% - i.e. the servers).Does not mean all code has to be duplicated but the inner loops where the 80%/20% rule (i.e. 20% of the code that consumes 80% of the time, which in my experience often becomes like the 99%/1% rule) comes into play that's where you write two code paths:
1 for AVX-512 in case it CPU is capable and
2 with just AVX2 in case CPU is not capable
mostly this ends up being just as I said the inner most loops, and there are excellent broadly available templates to use for this.
Just from a pure comp sci perspective it is quite interesting to vectorize code and see the benefits - pretty cool actually.