4 ms·
same ballpark as a 105w ryzen 5900x desktop chip with slight differences in single core/multi core. I'm impressed anyhow.
by eliasmacpherson 5y ago
same ballpark as a 105w ryzen 5900x desktop chip with slight differences in single core/multi core.
I'm impressed anyhow.
- DeathArrow 5y agoI would trust more other benchmarks like SPEC or Cinebench. However, even in Geekbench, 5950X reach 21k multicore score and 2300 single core.
- eliasmacpherson 5y agoThanks for the info, I have not paid much attention to the 5950x!
- GrayShade 5y agoDamn, I get 1736 / 17985 with PBO, but without overclocking otherwise. But I wouldn't trust Geekbench across platforms.
- adrian_b 5y agoIf you get only 1736 single-core, then you have probably run the benchmark on Windows. The scores of many benchmarks, especially the scores of single-thread benchmarks, are lower and also much more variable on Windows than either on Linux or on MacOS. My assumption is that this is due to the fact that almost all Windows computers run continuously many junk programs, e.g. antiviruses and the like, in parallel with whatever you are doing, and those interfere with the benchmarks. Running the same benchmark on Linux or maybe on a Windows completely cleaned of garbage and with the useless services stopped, should give results well above 1800 in non-overclocked single-thread.
- PaulKeeble 5y agoIts mostly due to the way Windows switches threads between cores but also scheduling and OS overhead. You can get a lot closer to Linux if you use software to pin the benchmarking threads to a particular core. The tweaks they did for AMD stop the switching moving from the CCX so often but they don't fix the detrimental effect on cache performance of switching threads to different cores ever 10ms.
- fmajid 5y agoAlso Intel's Clear Linux usually yields a big improvement over Ubuntu. Not relevant to ARM64, of course.
- GrayShade 5y agoNo, I actually ran it on Arch, but I don't exclude some variation caused by the kernel configuration.
- adrian_b 5y agoIt is also possible that the benchmark was run on one of the slow cores. Ryzen 9 5900X or 5950X will have a few fast cores, maybe 2 fast cores, which will run a ST task at 4.8 GHz, respectively at 4.9 GHz. If you are lucky, and more likely if you use PBO, the fast cores of your CPU might run at a slightly higher frequency, e.g. by 100 MHz, than guaranteed, especially when you have a good cooler that keeps the CPU at a low temperature. However the rest of the cores will have a lower maximum frequency than the nominal frequency, e.g by 100 MHz or by 200 MHz. The Linux scheduler, at least the default one, does not schedule the threads preferentially on the fast cores, so when running a ST benchmark there are more chances to run it on a slow core than on a fast core. You can determine manually which are your fast cores, by running some single-threaded benchmark pinned successively on each core, e.g. using taskset, while monitoring the clock frequency, e.g. with turbostat. The 1736 value looks like a normal value for a slow core. Rerunning the benchmark, but pinned on a fast core with taskset, should give on a 5950X a value not far from 1850. I have a slower 5900X and with a memory slower than normal (only 2667) and no PBO, but I still get over 1800 on a fast core.
- fmajid 5y agoThere are a lot of aberrant Geekbench scores, their results are not authenticated.
- adrian_b 5y ago2300 single core must be for an overclocked 5950x. Single-core scores around 1850 are normal for a 5950X with default settings, i.e. no overclocking. That is better than Apple M1, but not by much.
- tromp 5y agoWhat is the TDP of the M1 Max? Presumably a lot lower than 100W as it runs in a laptop.
- pja 5y agoAnandtech measured the TDP of the M1 at ~28W & this chip has twice the number of high power cores & 2 fewer low power cores, so probably somewhere around 40-45W? Eventually the rest of the market will get access to TSMC’s 5nm capacity & we’ll see what they can do on the same process.