Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AMD Threadripper 2990WX 32-Core and 2950X 16-Core Review

anandtech.com

Post reply on HN
Print view View original

131–140 of 202 posts

Re: AMD Threadripper 2990WX 32-Core and 2950X 16-Core Review

#131

Post by aurailious »

sounds wrote: post #118
blattimwind wrote: post #12

> After core counts, the next battle will be on the interconnect. Low power, scalable, and high performance: process node scaling will mean nothing if the interconnect becomes 90% of the total chip power. I think this means that scaling processors to more cores (nodes) in a lightweight-NUMA scheme just isn't going to work. My guess going forward is that the core counts will stay at the current level for quite a while…

I was hoping they would at least make a passing reference to Amdahl's Law. As you optimize one part of the system so that it is blazingly fast or lower power or whatever, then the amount taken up by the rest of the system increases proportionally. (In other words, real performance wins happen gradually as all the parts receive their own little optimizations which add up.)

Sounds a lot like how to play factorio.

Re: AMD Threadripper 2990WX 32-Core and 2950X 16-Core Review

#132

Post by xzel »

derekdahmer wrote: post #98
xzel wrote: post #31

Earlier quoted context omitted.

These recentish AMD core improvements alone have made me considering rebuilding my VST collection for Windows and moving off mac for production. I'm not looking forward to tracking down windows VST versions of tiny apps. I can't imagine spending 3k for another Macbook when I can get way more interesting performance in Windows these days. I'd love to have a desktop for very heavy synth and processing work, freeze thos…

I was in a similar position, really tied to all my tools on OS X but didn't want to spent 3k on their desktop line. Instead I built an awesome Hackintosh with 16gb ram, 8 cores, nvme drive and 1080 for like $1600 that runs High Sierra. It's definitely more work to set up initially but pretty low hassle afterwards. No regrets.

Yes I'm at the crossroads but I want to be able to have a laptop to take as well and the new mac line up is just not for me. Thats great info that the Hackintosh computers are still kicking because I had mostly ruled that out after not hearing much about them over the past few years. Mind sharing your build? :D

Re: AMD Threadripper 2990WX 32-Core and 2950X 16-Core Review

#133

Post by celrod »

deng wrote: post #54

This is the perfect chip for continuous, low-memory number crunching. For everything else... not so much. I mean, this chip consumes 74W when idle, of which almost 90% are spend on the interconnect. That's insane. Most important bit in the review for me: "After core counts, the next battle will be on the interconnect. Low power, scalable, and high performance: process node scaling will mean nothing if the interconnec…

If that number crunching is highly vectorizable, avx-512 of skylake-x will likely still be faster. Eg, an Intel 7900x is over twice as fast at multiplying big matrices (multithreaded) as an AMD 1950x. Of course, a similarly priced GPU would stomp both whenever CPU-GPU latency isn't much of an issue.

Meaning "highly vectorizable number crunching that you can't practically run on a GPU for some reason" is probably fairly niche. In LINPACK the even the 1950x beat the 7980XE on phoronix: https://www.phoronix.com/scan.php?page=article&item=amd-linu... and I'd have thought LINPACK would really reward AVX...

Maybe it isn't optimized to take advantage of it? I love wasting time playing with number crunchinng microbenchmarks, and when you do that avx-512 seems so unassailably impressive, it's surprising to see this not bourne out in larger benchmarks. (Also worth pointing out you need to tell gcc -mprefer-vector-width=512 on top of march=skylake-avx512, otherwise it'll prefer 256 bit vectors).

Anyway, those benchmarks definitely support compiling and number crunching. Pretty ideal for stuff like monte carlo.

Re: AMD Threadripper 2990WX 32-Core and 2950X 16-Core Review

#134

Post by Twirrim »

sandworm101 wrote: post #124
ReverseCold wrote: post #70

Earlier quoted context omitted.

KSP 1.1 finally added multithreading btw

When launching a rocket, that rocket's physics engine is still on a single thread. Given that the physics engine is the most CPU-intensive part of KSP, I'd say that KSP's support of multithreading is nominal at best. Multithreading aside: KSP's approach to physics also doesn't help them. Rather than make creative decisions re what should and shouldn't be calculated, they just hammer the physics thread with lots of ca…

KSP is now owned by Take Two Interactive (since May). Here's hoping they can get those dev resources!

Re: AMD Threadripper 2990WX 32-Core and 2950X 16-Core Review

#136

Post by dragontamer »

karangoeluw wrote: post #2

Honest q: What home-use workstation would use 32 cores? (Excluding home labs or servers).

It takes roughly 20-minutes to make a SINGLE good frame using Cycles on Blender. Cycles is a raytracer for 3d modeling.

If you are making a 30-second animation at 24-frames-per-second, that would be 720 frames, or roughly 240 hours (10 days) of rendering. 30-seconds would be roughly the length of a standard commercial.

If you have a computer that is 2x or 4x faster, that cuts the time down to 5 days or 2.5 days. Which is dramatically different. Its mostly a CPU-intensive problem with relatively low RAM bandwidth. (Its RAM-heavy, especially with HDR skymaps. So you need lots of RAM but not necessarily fast RAM).

Re: AMD Threadripper 2990WX 32-Core and 2950X 16-Core Review

#137

Post by dragontamer »

somebodythere wrote: post #103
c2h5oh wrote: post #51

Earlier quoted context omitted.

AMD has manufacturing cost advantage - CPUs are built from 4- core CCXes that can be binned separately. Intel on the other hand needs to manufacture a monolithic CPU that not only is fault free in enough cores, but performs well. That's harder and yields are way lower. 80% yield on a 4 core block is a 16.7% yield on a 32 core block - and that's before binning

If AMD's method scales so much easier, why didn't Intel just... do that? Honest question.

Intel could do that. But AMD came out with this method first.

AMD has only been doing this "infinity fabric" thing for a year. Intel was caught with their pants down. It seems like Intel is researching chiplet technology and trying to recreate AMD's success here.

It takes several years to create chips. So Intel realistically won't be able to copy the strategy until 2020 or later. But you better bet that Intel is going to be investing heavily into chiplet technology, now that AMD demonstrated how successful it can be.

Re: AMD Threadripper 2990WX 32-Core and 2950X 16-Core Review

#139

Post by dragontamer »

florabuzzword wrote: post #76
karangoeluw wrote: post #2

Honest q: What home-use workstation would use 32 cores? (Excluding home labs or servers).

Does video editing or After Effects work could as home use yet? Video editing, in terms of cutting and splicing clips, is not going to benefit much from this chip, but a lot of effects rendering will benefit, and basic video editing uses a lot of effects these days. That’s all I’ve got.

Yes for roughly 8-threads or so.

16-thread Threadripper is mostly idle in my video-editing tests. I mean, its a great processor. But video editing isn't "heavy enough" for me to recommend a 16-core or bigger processor.

Re: AMD Threadripper 2990WX 32-Core and 2950X 16-Core Review

#140

Post by greglindahl »

somebodythere wrote: post #103
c2h5oh wrote: post #51

Earlier quoted context omitted.

AMD has manufacturing cost advantage - CPUs are built from 4- core CCXes that can be binned separately. Intel on the other hand needs to manufacture a monolithic CPU that not only is fault free in enough cores, but performs well. That's harder and yields are way lower. 80% yield on a 4 core block is a 16.7% yield on a 32 core block - and that's before binning

If AMD's method scales so much easier, why didn't Intel just... do that? Honest question.

Intel has done that in the past, actually, their first "dual core" chip (2005) was actually two chips in a package.

https://en.wikipedia.org/wiki/Pentium_D

Post reply on HN