AMD Rome Second Generation EPYC Review: 2x 64-core Benchmarked
by Johan De Gelas on August 7, 2019 7:00 PM ESTSingle-Thread SPEC CPU2006 Estimates
While it may have been superceded by SPEC2017, we have built up a lot of experience with SPEC CPU2006. Considering the trouble we experience with our datacenter infrastructure, it was our best first round option for raw performance analysis.
Single threaded performance continues to be very important, especially in maintainance and setup situations. These examples may include running a massive bash script, trying out a very complex SQL query, or configuring new software - there are lots of times where a user simply does not use all the cores.
Even though SPEC CPU2006 is more HPC and workstation oriented, it contains a good variety of integer workloads. It is our conviction that we should try to mimic how performance critical software is compiled instead of trying to achieve the highest scores. To that end, we:
- use 64 bit gcc : by far the most used compiler on linux for integer workloads, good all round compiler that does not try to "break" benchmarks (libquantum...) or favor a certain architecture
- use gcc version 7.4 and 8.3: standard compiler with Ubuntu 18.04 LTS and 19.04.
- use -Ofast -fno-strict-aliasing optimization: a good balance between performance and keeping things simple
- added "-std=gnu89" to the portability settings to resolve the issue that some tests will not compile
The ultimate objective is to measure performance in non-aggressively optimized"applications where for some reason – as is frequently the case – a multi-thread unfriendly task keeps us waiting. The disadvantage is there are still quite a few situations where gcc generates suboptimal code, which causes quite a stir when compared to ICC or AOCC results that are optimized to look for specific optimizations in SPEC code.
First the single threaded results. It is important to note that thanks to turbo technology, all CPUs will run at higher clock speeds than their base clock speed.
- The Xeon E5-2699 v4 ("Broadwell") is capable of boosting up to 3.6 GHz. Note: these are old results compiled w GCC 5.4
- The Xeon 8176 ("Skylake-SP") is capable of boosting up to 3.8 GHz.
- The EPYC 7601 ("Naples") is capable of boosting up to 3.2 GHz.
- The EPYC 7742 ("Rome") boosts to 3.4 GHz. Results are compiled with GCC 7.4 and 8.3
Unfortunately we could not test the Intel Xeon 8280 in time for this data. However, the Intel Xeon 8280 will deliver very similar results, the main difference being that it runs a 5% higher clock (4 GHz vs 3.8 GHz). So we basically expect the results to be 3-5% higher than the Xeon 8176.
As per SPEC licensing rules, as these results have not been officially submitted to the SPEC database, we have to declare them as Estimated Results.
Subtest | Application Type | Xeon E5-2699 v4 |
EPYC 7601 |
Xeon 8176 |
EPYC 7742 |
EPYC 7742 |
Frequency | 3.6 GHz | 3.2 GHz | 3.8 GHz | 3.4 GHz | 3.4 GHz | |
Compiler | gcc 5.4 | gcc 7.4 | gcc 7.4 | gcc 7.4 | gcc 8.3 | |
400.perlbench | Spam filter | 43.4 | 31.1 | 46.4 | 41.3 | 43.7 |
401.bzip2 | Compression | 23.9 | 24.0 | 27.0 | 26.7 | 27.2 |
403.gcc | Compiling | 23.7 | 35.1 | 31.0 | 42.3 | 42.6 |
429.mcf | Vehicle scheduling | 44.6 | 40.1 | 40.6 | 39.5 | 39.6 |
445.gobmk | Game AI | 28.7 | 24.3 | 27.7 | 32.8 | 32.7 |
456.hmmer | Protein seq. | 32.3 | 27.9 | 35.6 | 30.3 | 60.5 |
458.sjeng | Chess | 33.0 | 23.8 | 32.8 | 27.7 | 27.6 |
462.libquantum | Quantum sim | 97.3 | 69.2 | 86.4 | 72.7 | 72.3 |
464.h264ref | Video encoding | 58.0 | 50.3 | 64.7 | 62.2 | 60.4 |
471.omnetpp | Network sim | 44.5 | 23.0 | 37.9 | 23.0 | 23.0 |
473.astar | Pathfinding | 26.1 | 19.5 | 24.7 | 25.4 | 25.4 |
483.xalancbmk | XML processing | 64.9 | 35.4 | 63.7 | 48.0 | 47.8 |
A SPEC CPU analysis is always complicated, being a mix of what kind of code the compiler produces and CPU architecture.
Subtest | Application type | EPYC 7742 (2nd gen) vs 7601 (1st gen) |
EPYC 7742 vs Intel Xeon Scalable |
Gcc 8.3 |
400.perlbench | Spam filter | +33% | -11% | +6% |
401.bzip2 | Compression | +11% | -1% | +2% |
403.gcc | Compiling | +21% | +28% | +1% |
429.mcf | Vehicle scheduling | -1% | -3% | 0% |
445.gobmk | Game AI | +35% | +18% | +0% |
456.hmmer | Protein seq. analyses | +9% | -15% | +100% |
458.sjeng | Chess | +16% | -16% | -1% |
462.libquantum | Quantum sim | +5% | -16% | -1% |
464.h264ref | Video encoding | +24% | -4% | -3% |
471.omnetpp | Network sim | +0% | -39% | 0% |
473.astar | Pathfinding | +30% | +3% | 0% |
483.xalancbmk | XML processing | +36% | -25% | 0% |
First of all, the most interesting datapoint was the fact that the code generated by gcc 8 seems to have improved vastly for the EPYC processors. We repeated the single threaded test three times, and the rate numbers show the same thing: it is very consistent.
hmmer is one of the more branch intensive benchmarks, and the other two workloads where the impact of branch prediction is higher (somewhat higher percentage of branch misses) - gobmk, sjeng - perform consistingly better on the second generation EPYC with it's new TAGE predictor.
Why the low IPC omnetpp ("network sim") does not show any improvement is a mystery to us, we expected that the larger L3 cache would help. However this is a test that loves very large caches, as a result the Intel Xeons have the advantage (38.5 - 55 MB L3).
The video encoding benchmark "h264ref" also relies somewhat on the L3 cache, but that benchmark relies much more on DRAM bandwidth. The fact that the EPYC 7002 has higher DRAM bandwidth is clearly visible.
The pointer chasing benchmarks – XML procesing and Path finding – performed less than optimal on the previous EPYC generation (compared to the Xeons), but show very significant improvements on EPYC 7002.
180 Comments
View All Comments
krumme - Thursday, August 8, 2019 - link
Because he is feeded by another hand.Enjoy the objectivity by Johan as its is very rare these days. It's not easy for AT to post this stuff. So kudos to them.
hoohoo - Thursday, August 8, 2019 - link
Nice review, but tbh I think you should run the AMD system as such, not limit it's RAM to what the Intel system maxes out at. I would not buy a system and configure it to limits of the competition: I would configure it to it's actual linits.yankeeDDL - Thursday, August 8, 2019 - link
Wow. "Blasted" is the only word that comes to mind. Good job AMD.eastcoast_pete - Thursday, August 8, 2019 - link
Thanks Johan and Ian! Impressive results, glad to see that AMD is once again making Intel sweat, all of which can only be good for us.Question: A bit out of left field, but why does AMD put the 7 nm dies in these close pairs, as opposed to leaving a little more space between them? Wouldn't thermals be better if each chip gets a little more "reserved" lid space? Just curious. Thanks!
sharath.naik - Thursday, August 8, 2019 - link
Now since we finally are entering the era where a single server(Yes backup is addition) is enough for most of smaller organizations. There is one thing that is needed, OS limits/zones which can limit the cpus and memory built in, instead of using VMs. This will save a lot on resources wasted on booting up an entire OS for individual applications. Linux has the ability for targeting specific cpus but not sure windows has it. But there is a need for standardized way to limit resources by process and by user.mdriftmeyer - Thursday, August 8, 2019 - link
Agreed.quorm - Thursday, August 8, 2019 - link
Is it possible you haven't heard of docker?abufrejoval - Sunday, August 11, 2019 - link
or OpenVZ/Virtuozzo or quite simply cgroups. Can even nest them, including with VMs.DillholeMcRib - Thursday, August 8, 2019 - link
destruction … Intel sat on their proverbial hands too long. It's over.crotach - Friday, August 9, 2019 - link
Bye Intel!!