forums.silverfrost.com

[JohnCampbell]

Dan,

I bought a Ryzen 5900X last Dec-20 and it is much faster than my previous i7-8700K. Actually it’s between 50% to 100% faster for my FE analysis, depending on the type of calculation.
I first got a 5900X + 64GB 3600MHz memory, but it kept crashing on multi-thread calcs. Changed to 3200MHz memory and it now doesn't crash. Presumably the quality of the silicon in the 3600MHz memory was a problem. I am not sure of the silicon quality of the 5900X !!

For my large array calcs using more threads, but only 2 memory channels is a significant bottleneck. I don't get much better performance above threads = cores, (which is similar with the 8700K)

I tried to find an easy problem to define and apply OpenMP! I have been doing testing of large matrix multiply using my developed code:
C[15000,12000] = A[15000,11000] x B[11000,12000] (see equation.com),
where partitioning is essential to reduce the memory<>cache bottleneck. (Vectors must be in cache for AVX to work efficiently and there are 3 levels of cache!) My main measure of performance is to calculate the number of floating point multiplies per second, as GFLOPS (10^9 flop/Sec). My coding approaches at partitioning produce 50 Gflop/s for i7-8700 and 100 Gflop/s for 5900X. These are significantly slower than MKL - DGEMM claimed performance (250+ Gflop/s for similar i5 processors), that I cannot approach (even allowing for MKL benchmarks count additions) (Equation.com report 22 Gflop/s for Opteron and 9.7 Gflop/s for Xeon which is slow)

Interesting that the Ryzen shows significant variability in gflops vs threads for my coding approaches, especially as threads exceeds cores. I7-8700 similarly stalls as threads exceed cores. This is an area I need to investigate further. My next processor will have more memory channels.
OpenMP with large arrays is not an easy coding problem. (large is array size >> cache size)
I will try to post some results when I can better describe the problem.
You can't just buy a different processor and use it. There is lots of tuning to do.

[DanRRight]

John,
So in summary you have got twice more cores inside Ryzen and 50 to 100% increase vs Intel ? Does this mean that the Ryzen single core performance is around the same as with Intel ?

Unfortunately i do not have anyone nearby with larger memory channel PCs. I have access to 10000 core Linux supercomputer which uses older Intel 12 core Xeon processors which would be not so interesting to test, and the code we use is written in C. Fortran version 19 with AVX should be there too but there is no one to ask how to use it, the good sysadmin left the team.

The only person i know by contacting him few years ago who has broad access to all world existing processors and who is also interested to test them is Ian Cutress from Anandtech. The UK guy by the way, former scientist, nice and easy going person, at least he was in the past before he started interviewing all the top CEOs in the IT industry. Try to convince him to run the test on 4, 6 and 8 memory channel computers. His own 3D particle moving code got huge benefits from AVX512. Plus he knew the former engineer at Intel who adjusted his code with AVX to get 3-4 even 5x increase in performance vs no-AVX. If he will find that some processors favor significantly cache size, memory channels or AVX with such important task as linear algebra i am sure there will be huge buzz in the industry. He touted his AVX speed increase with Intel processors vs AMD which do not have AVX512 last few years, and Intel clearly liked this. When we implemented in our codes AVX512 though the increase in performance was just 20% or less.

[JohnCampbell]

[DanRRight]

Noctua is good air cooler, one of the best, but i still recommend to use reliable good company water cooler.

[John-Silver]

Being an FE 'oldie' I'd be very interested if the vendors got their act together and came up with some reliable software (reliable in the sense of ALWAYS giving a speed increase) and of the right order of magnitude (that's orDER OF MAGNITUDE i.e. multiple(s)of ten.
Only then could any self-respecting organization pay a significant premium for performance improvement. Maybe they already have it (the improvement) but are continually playing 'silly bugges' trying to milk the market for all it's worth wit all that 'double the speed increase every 2 years crap ? ... as in the past
Aat the end of the day they have a further problem - FE speed gains are fine, UP TO A LIMIt, because as any self-respecting structural analyst knows, just simply increasing the size of a moe=del is no solution, since details in models oft introduce more problems than any speed increase can solve, getting wronger answers more quickly isn't a solution
_________________
''Computers (HAL and MARVIN excepted) are incredibly rigid. They question nothing. Especially input data.Human beings are incredibly trusting of computers and don't check input data. Together cocking up even the simplest calculation ... "

[John-Silver]

I'd be interested to know what size FE problems eople are dealing with out there which are 'driving' these performance 'battles', because inho a hundred thousand dofshould be enough to solve any problem !!!
_________________
''Computers (HAL and MARVIN excepted) are incredibly rigid. They question nothing. Especially input data.Human beings are incredibly trusting of computers and don't check input data. Together cocking up even the simplest calculation ... "

[LitusSaxonicum]

JC,

Is that a self-build, or a commercial pre-built system? If you built it, what case and fans did you use? A system built into a tower case shouldn't have thermal throttling.

Eddie

[DanRRight]

Also, i suggest to find on the internet some PC sellers with Threadrippers and ask them to run your benchmark. For example 3960x has 2x more cores, 2x more cache, 2x more memory channels. Processor also costs 2x vs 5900x and consumes at peak 2.5x more but still the whole PC will cost probably just 30-50% more if build it by yourself. There are a lot of testers on the internet who might be interested. Threadripper Pro is also coming in few months (i do not see first samples of it are much faster though)

On the net I find some insanely expensive prebuilt workstations for $8k with Threadrippers, i would make them myself for 2-3x less

[John-Silver]

from wot i've seen in posts on here over the last 5 years or so, multi-tasking multi-cores have a very long long way to go before they become interesting?
Anarea where the PC manufacturers hav (so far) fallen down imho.
So much hype and so little result.
So many people duped into buying ott non-what's written on the packet hurdy gurdys.

maybe we should all revert to FTN77 with machines to match.
_________________
''Computers (HAL and MARVIN excepted) are incredibly rigid. They question nothing. Especially input data.Human beings are incredibly trusting of computers and don't check input data. Together cocking up even the simplest calculation ... "

[John-Silver]

... even JohnC and his FE codes !!!
Who needs any model greater than 10000 ndes JohnC when we usd to use 3000 node models with impunity in 1980 when I started work !!! LOL

(in 1997ish I saw a pretty large organisation brought to its knees one sunny Friday aftenoon by an engineer-ette who did the rounds of all the offices asking EVERYONE to kill their NASTRAN jobs or hers would crash - it was an electronic box of something like 100000 nodes , which was mega-enormous at the time !!!

We aint got any better almost a quarterof a century later !
_________________
''Computers (HAL and MARVIN excepted) are incredibly rigid. They question nothing. Especially input data.Human beings are incredibly trusting of computers and don't check input data. Together cocking up even the simplest calculation ... "

[JohnCampbell]