AMD Ryzen 7 1800X vs. Intel Core i7 7700K Linux Gaming Performance

[zboson]

I can control cores, I will have tests up this weekend.

Michael, I have not read all the comments here but if you have not agreed to already can you redo your tests with SMT disabled? This should remove any bias due to thread topology. Is there anyway two confirm which threads are going to which cores?

Intel calls their SMT implementation Hyper-Threading. What does AMD call their implementation of SMT (bulldozer had cluster multi-threading not SMT)? Anyway, that's what I mean. Disable AMD's equivalence of Hyper-Threading.

[Evontroy]

Can you submit your CSGO and Dota 2 results to Valve please.

[DRanged]

But beyond that, the BIOS version isn't exposed (maybe via dmidecode when running as root, haven't checked in this particular instance) and thus isn't something PTS can parse and report to the benchmarking results. If there happens to be some interface available to reliably read a board's BIOS version, happy to add support to PTS for it.

Running biosdecode (included in dmidecode) | grep SMBIOS as root gives you the bios revision, for your purposes it needs to finetuned a bit .

Or have a look into cpu-x, it has cli and graphical interface.

[cRaZy-bisCuiT]

Welcome to Phoronix Fanboyism - why would you compare those CPUs with stock clock? Won't the 1800x be capable of 4-4,2 Ghz? The IPC has increased a lot, still it can't compete to the 7700k clock speed when it differs with more than 1 Ghz .... at least when it boils down to IPC or the performance of 1-4 cores.

[Michael]

Welcome to Phoronix Fanboyism - why would you compare those CPUs with stock clock? Won't the 1800x be capable of 4-4,2 Ghz? The IPC has increased a lot, still it can't compete to the 7700k clock speed when it differs with more than 1 Ghz .... at least when it boils down to IPC or the performance of 1-4 cores.

??? Because almost always the tests are done at stock speeds to show out-of-the-box performance, since that's what you get when you buy said product? But as said in the earlier article, I will deliver some Ryzen overclocking results soon.

[coastiron]

can you make your test again, with Nvidia GPUs?
TY!

I also think that they would be worth doing. Nvidia GPUs and proprietary drivers.
Don't get me wrong: I am big enthusiast of open-source drivers, but at launch day of completely new CPU microarchitecture the are usually quite not optimized.

[torsionbar28]

Welcome to Phoronix Fanboyism - why would you compare those CPUs with stock clock?

Because overclockers are a miniscule fraction of the market. Most people want to buy a product and have it perform as intended without tinkering with it. Overclocking is for hobbyists living in their parents basement. Professional users aren't interested in overclocking whatsoever.

[Michael_S]

Good point. I wonder if the performance difference being seen is related to the thread topology. Last time I looked into thread topology of AMD bulldozer based processor the default topology I noticed on Linux and Windows was compact. In other words if you asked for two threads it would pack two threads to one module. That's normally going to much less efficient than putting two threads on two different modules (especially as bulldozer based processors only have one FPU per module). However with Intel hyper threading on each Linux system I have tested the topology is scatter. Each thread goes to a separate core until all cores have a thread and only after can two threads go to the same core.

So what could be happing with Ryzen is that an application/game asks for two threads and those go to one core or four threads which only go to two core, eight threads to four cores. That could account for a factor of 2 difference with Intel.

OpenMP by default (at least on all the systems I have tested it on) will use all logical processors. So on Ryzen it would use 16 threads. That could explain why OpenMP based rendering applications perform so well. If you use the maximum threads (16) then there is no problem with topology but if you use less than 16 threads the topology can have a big effect on performance.

Edit: I just checked cpuinfo and I think it's compact for Ryzen but scattered for Intel

On my Skylake 4C/8HT system with Ubuntu 16.10

grep "core id" /proc/cpuinfo returns

core id : 0
core id : 1
core id : 2
core id : 3
core id : 0
core id : 1
core id : 2
core id : 3

But the same command with the 1800X returns
http://openbenchmarking.org/system/1...01800X/cpuinfo

core id : 0
core id : 0
core id : 1
core id : 1
core id : 2
core id : 2
core id : 3
core id : 3
core id : 4
core id : 4
core id : 5
core id : 5
core id : 6
core id : 6
core id : 7
core id : 7

That's a great point about the thread splitting. But it would be bizarre for AMD not to have addressed this right out of the gate, if it's a common issue. "We're going to spend billions of dollars designing Bulldozer and later Ryzen, but not spend a few million getting the necessary patches into the Linux kernel and Windows kernel for the multi-threading to work efficiently." ??

I believe it's possible that in some rare cases if a single application is splitting threads across a single core it might be better than splitting across multiple cores. If there is frequent intra-thread communication, the shared L2 cache could allow that communication to run much faster. But that's my wild speculation.

[dungeon]

Welcome to Phoronix Fanboyism - why would you compare those CPUs with stock clock? Won't the 1800x be capable of 4-4,2 Ghz? The IPC has increased a lot, still it can't compete to the 7700k clock speed when it differs with more than 1 Ghz .... at least when it boils down to IPC or the performance of 1-4 cores.

Maybe because beside Radeon Software's and Wattman, we are now also missing easy to breaking things Ryzen Master app

Hmmm... or maybe it is other way for Linux, so we are counting on being safe or having virtually less bugs because console is future or something

Just in case, here is the AMD's official OC warning, that no one actually read until it is too late

WARNING: AMD processors, including chipsets, CPUs, APUs and GPUs (collectively and individually “AMD processor”), are intended to be operated only within their associated specifications and factory settings. Operating your AMD processor outside of official AMD specifications or outside of factory settings, including but not limited to the conducting of overclocking (including use of this overclocking software, even if such software has been directly or indirectly provided by AMD or an entity otherwise affiliated in any way with AMD), may damage your processor, affect the operation of your processor or the security features therein and/or lead to other problems, including but not limited to damage to your system components (including your motherboard and components thereon (e.g., memory)), system instabilities (e.g., data loss and corrupted images), reduction in system performance, shortened processor, system component and/or system life, and in extreme cases, total system failure. It is recommended that you save any important data before using the tool. AMD does not provide support or service for issues or damages related to use of an AMD processor outside of official AMD specifications or outside of factory settings. You may also not receive support or service from your board or system manufacturer. Please make sure you have saved all important data before using this overclocking software. DAMAGES CAUSED BY USE OF YOUR AMD PROCESSOR OUTSIDE OF OFFICIAL AMD SPECIFICATIONS OR OUTSIDE OF FACTORY SETTINGS ARE NOT COVERED UNDER ANY AMD PRODUCT WARRANTY AND MAY NOT BE COVERED BY YOUR BOARD OR SYSTEM MANUFACTURER’S WARRANTY.

[Michael_S]

Because overclockers are a miniscule fraction of the market. Most people want to buy a product and have it perform as intended without tinkering with it. Overclocking is for hobbyists living in their parents basement. Professional users aren't interested in overclocking whatsoever.

I agree with half. I agree that overclockers are a tiny fraction of the market, and it makes the most sense to run benchmarks at stock speeds.

I disagree that overclocking is for people living in their parents' basement. I never tried overclocking at 20 or 30 because if I screwed up and bricked my CPU it was a major hit to my budget. Now I'm 40, most of my student loans are paid off, most of my kids are out of daycare, and my career is probably at its peak. (The next time I go job hunting my gut and grey hair will probably be a big problem. Yay technology industry. But I digress.) Today I wouldn't be thrilled if I roasted a $500 CPU, but I could order a replacement the same day without blinking an eye.