AMD Ryzen 9 5980HS Cezanne Review: Ryzen 5000 Mobile Tested
by Dr. Ian Cutress on January 26, 2021 9:00 AM EST- Posted in
- CPUs
- AMD
- Vega
- Ryzen
- Zen 3
- Renoir
- Notebook
- Ryzen 9 5980HS
- Ryzen 5000 Mobile
- Cezanne
CPU Tests: Synthetic and SPEC
Most of the people in our industry have a love/hate relationship when it comes to synthetic tests. On the one hand, they’re often good for quick summaries of performance and are easy to use, but most of the time the tests aren’t related to any real software. Synthetic tests are often very good at burrowing down to a specific set of instructions and maximizing the performance out of those. Due to requests from a number of our readers, we have the following synthetic tests.
Linux OpenSSL Speed: SHA256
One of our readers reached out in early 2020 and stated that he was interested in looking at OpenSSL hashing rates in Linux. Luckily OpenSSL in Linux has a function called ‘speed’ that allows the user to determine how fast the system is for any given hashing algorithm, as well as signing and verifying messages.
OpenSSL offers a lot of algorithms to choose from, and based on a quick Twitter poll, we narrowed it down to the following:
- rsa2048 sign and rsa2048 verify
- sha256 at 8K block size
- md5 at 8K block size
For each of these tests, we run them in single thread and multithreaded mode. All the graphs are in our benchmark database, Bench, and we use the sha256 results in published reviews.
GeekBench 5: Link
As a common tool for cross-platform testing between mobile, PC, and Mac, GeekBench is an ultimate exercise in synthetic testing across a range of algorithms looking for peak throughput. Tests include encryption, compression, fast Fourier transform, memory operations, n-body physics, matrix operations, histogram manipulation, and HTML parsing.
Unfortunately we are not going to include the Intel GB5 results in this review, although you can find them inside our benchmark database. The reason behind this is down to AVX512 acceleration of GB5's AES test - this causes a substantial performance difference in single threaded workloads that thus sub-test completely skews any of Intel's results to the point of literal absurdity. AES is not that important of a real-world workload, so the fact that it obscures the rest of GB5's subtests makes overall score comparisons to Intel CPUs with AVX512 installed irrelevant to draw any conclusions. This is also important for future comparisons of Intel CPUs, such as Rocket Lake, which will have AVX512 installed. Users should ask to see the sub-test scores, or a version of GB5 where the AES test is removed.
To clarify the point on AES. The Core i9-10900K scores 1878 in the AES test, while 1185G7 scores 4149. While we're not necessarily against the use of accelerators especially given that the future is going to be based on how many and how efficient these accelerators work (we can argue whether AVX-512 is efficient compared to dedicated silicon), the issue stems from a combi-test like GeekBench in which it condenses several different (around 20) tests into a single number from which conclusions are meant to be drawn. If one test gets accelerated enough to skew the end result, then rather than being a representation of a set of tests, that one single test becomes the conclusion at the behest of the others, and it's at that point the test should be removed and put on its own. GeekBench 4 had memory tests that were removed for Geekbench 5 for similar reasons, and should there be a sixth GeekBench iteraction, our recommendation is that the cryptography is removed for similar reasons. There are 100s of cryptography algorithms to optimize for, but in the event where a popular tests focuses on a single algorithm, that then becomes an optimization target and becomes meaningless when the broader ecosystem overwhelmingly uses other cryptography algorithms.
CPU Tests: SPEC
SPEC2017 and SPEC2006 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.
We run the tests in a harness built through Windows Subsystem for Linux, developed by our own Andrei Frumusanu. WSL has some odd quirks, with one test not running due to a WSL fixed stack size, but for like-for-like testing is good enough. SPEC2006 is deprecated in favor of 2017, but remains an interesting comparison point in our data. Because our scores aren’t official submissions, as per SPEC guidelines we have to declare them as internal estimates from our part.
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
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 labelled ‘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 each of the SPEC targets we are doing, SPEC2006 rate-1, SPEC2017 speed-1, and SPEC2017 speed-N, rather than publish all the separate test data in our reviews, we are going to condense it down into a few interesting data points. The full per-test values are in our benchmark database.
Facebook
Twitter
RSS
218 Comments
View All Comments
Meteor2 - Thursday, February 4, 2021 - link
Great point.ikjadoon - Tuesday, January 26, 2021 - link
It's great to see AMD kicking Intel's butt in a much larger market (i.e., laptops vastly outsell desktops): AMD really should be alongside, or simply replacing, Intel in most premium notebooks. Gaming notebooks are not my cup of tea, but glad to see for upcoming 15W Zen3 parts.Will we see actual, high-end Zen3 notebooks? Lenovo, HP, ASUS, Dell: for shame if you keep ramming toasty Tiger Lake down customers' throats. Lenovo's done some great offerings with both AMD & Intel; that means some compromises with notebook design (just go all AMD, man; if/when Intel is on top, switch back!), but beefier cooling for Intel will also help AMD.
Still, overall, I don't see anything convincing me that x86 is really right for notebooks, either. So much waste heat...for what? The M1 has rightly rejiggered expectations: 20 hours on 150 nits should be ordinary, not miraculous. Limited to no fan spin-up and max CPU load should yield a chassis maximum of 40C (slightly warmer than body temperature). And, all the while with class-leading 1T performance.
As this is a gaming laptop, it's not too relevant to compare web benchmarks (what most laptops do), but this is peak Zen3 mobile and it still falls quite short:
Speedometer 2.0
35W Ryzen 5980HS: 102 points (-57%)
125W i9-10900K: 119 points (-49%)
35W i7-1185G7: 128 points (-46%)
105W Ryzen 5950X: 140 points (-40%)
30W Apple M1: 234 points
You can double / triple x86 wattage and still be miles behind M1. I almost feel silly buying an x86 laptop again: just kilowatts of waste heat over time. Why? Electrons that never get used, just exhausted and thrown out as soon as possible because it'll throttle even worse otherwise.
undervolted_dc - Tuesday, January 26, 2021 - link
because you here are benchmarking javascript engine in the browserbut not being enough you are comparing those in single thread so here you are comparing 1/16 of the 5950hs vs 1/4 of the m1
a 128core epyc or a 64core threadripper probably will be even worse in this single threaded benchmark ( because those are levaring threads and are less efficient in single threaded app )
if you like wrong calculations then 1 core of the 15w version use less tha 1w for what result ? ~ 100 points ? so who is wasting electrons here ?
( btw 1 core doesn't use 1/16 because there are boosts , but it's even less wrong than your comparison )
ZoZo - Tuesday, January 26, 2021 - link
128-core EPYC? Where?His comparison is indeed misleading in terms of energy efficiency, but it's sad that no x86 is able to come even close to that single-threaded performance.
WaltC - Tuesday, January 26, 2021 - link
Doubly sad for the M1 that we are living in the multicore/multithread era...;)ikjadoon - Tuesday, January 26, 2021 - link
The energy efficient comparisons are pretty clear: the best x86 (Zen3) has stunningly lower IPC than M1, which barely cracks 3 GHz. The only way to make up for such a gulf in IPC is faster clocks. Faster clocks require the 100+W TDPs so common in high-performance desktop CPUs. It's why Zen3 mobile clocks so much lower than Zen3 desktop (3-4 GHz instead of 4-5 GHz)A CPU that needs 3x power to do the same work (and do it slower in most cases) must exhaust an enormous amount of heat, when considering nT or 1T benchmarks (Zen3 requires ~20W for 5 GHz boost on a *single* core). Look at those boost power consumption measurements.
Specifically in desktops (noted in my comparison about tripling TDP...), the CPU *alone* eats up an extra 60 to 90 watts during peak usage. Call it +20W average continuously, so we can do the math.
20W x 8 hours x 7 days a week = +1.1 kWh excess exhaust heat per week. x86 had two corporate giants to do better. It's been severely litigated, but that's Intel's comeuppance. If Intel can't put out high-perf, high-efficiency x86 architectures, then people will start to feel less attached to x86 as an ISA. x86 had billions and billions and billions of R&D.
I see no reason for consumers to religiously follow x86 Wintel or Wintel-clones in laptops especially, but desktops, too: where is the efficiency going to be coming from? Even if Apple *had flat 1T* for the next three years, I'd still feel more optimistic about M1-based CPUs in the long-term than x86.
Dug - Tuesday, January 26, 2021 - link
"I see no reason for consumers to religiously follow x86 Wintel or Wintel-clones in laptops especially, but desktops, too: where is the efficiency going to be coming from?"Software, and getting work done. M1 is great and all, but just need to convince the boss that Apple or 3rd party has software available for our company....... Nope, oh well.
Other negatives-
For personal use, people aren't going to spend thousands of dollars to get new software on new platform.
They can't play games (or should I say they can't play a majority), which is probably the largest market.
They can't change anything about their software
They can't customize anything.
They can't upgrade any piece of their hardware.
They don't have options for same accessories.
So I'll go ahead and spend the extra $15 a year on energy to keep Windows.
Spunjji - Thursday, January 28, 2021 - link
"A CPU that needs 3x power to do the same work"It doesn't. It's been demonstrated a few times now that if you scale back Zen 3 cores to similar performance levels to M1, M1's perf/watt advantage drops to about 30%. It's still better than the node advantage alone, but it's not crippling, and M1 is simply not capable of scaling up to the clock speeds required to match x86 on desktop / HPC workloads.
They're different core designs matched to different purposes (ultra-mobile first vs. server first) and show different strengths as a result.
M1 is a significant achievement - no doubt about it - but you're *massively* overstating the case in its favour.
GeoffreyA - Friday, January 29, 2021 - link
Thank you for this.Meteor2 - Thursday, February 4, 2021 - link
"M1 is simply not capable of scaling up to the clock speeds required to match x86 on desktop / HPC workloads" ...Yet. In a couple of years x86 will be behind ARM across the board.Fastest HPC in the world is ARM *right now*. Only the fifth fastest is x86.