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

AMD Ryzen Threadripper 3000 32-core CPU is more bad news for Intel

zdnet.com

Post reply on HN
Print view View original

141–150 of 165 posts

Re: AMD Ryzen Threadripper 3000 32-core CPU is more bad news for Intel

#141

Post by paulmd »

loeg wrote: post #122
paulmd wrote: post #88

Earlier quoted context omitted.

2990WX was a weird ultra-NUMA setup where half the dies had no direct memory access at all. Part of the gain will have been fixing that - particularly on Windows where the scheduler did not understand this configuration at all (Linux results have always been much better). Another part is the power consumption and resulting higher clocks. The 2990WX was basically power limited, the new chips are on 7nm which will dras…

Wouldn't the higher clocks impact single-core performance? The article claims that's about the same. I suppose the userbenchmark single-core perf test[1] (the article links to some userbenchmark.com results) could just be a bad test; I'm unfamiliar with it. (It could also be that 2990WX wasn't power/heat limited on a single core, and neither is 3xxx; but say, at 16 core 100% clock, 3xxx can go higher than 2990WX due…

Zen2 is moving to a somewhat deceptive advertising model for clocks.

AMD is advertising the absolute highest that any core on the die can hit under extremely light load for an absolute instant. Sustained single-core clocks will be 100-200 MHz less and sustained all-core will be significantly less. Most cores on a chip are not capable of sustaining the advertised clockrate even under severe voltage and even for instant loads, only the "preferred core" on a chip. There is a significant "binning effect" not just between chips, but between individual cores on a chip.

Intel and previous generations of AMD cores used to advertise the sustained single-core rate, which could be achieved on any of the cores. This is a changeup to how the clockrate has been advertised.

original: https://www.youtube.com/watch?v=DgSoZAdk_E8

followup after a patch: https://www.youtube.com/watch?v=3LesYlfhv3o

As such the clockrates may be significantly different from what you're intuiting based on the advertising.

Re: AMD Ryzen Threadripper 3000 32-core CPU is more bad news for Intel

#142

Post by dragontamer »

d33 wrote: post #91
dragontamer wrote: post #83

Earlier quoted context omitted.

Dwarf Fortress is IIRC mainly limited by RAM latency. So you really should be trying to get 3600 MHz RAM with as tight timings (maybe CL16 timings?) as possible. Dwarf Fortress's simulation is all about pointer-indirection and jumping around memory. The CPU doesn't really do much except wait for RAM most of the time. It takes ~50ns to talk to RAM, but the CPU is clocked at 4GHz (0.25 nanoseconds), giving you an idea…

Interesting! Is there a simple way to measure memory latency?

Create 1-billion 32-bit integers numbers between 0 to 1-billion. Knuth shuffle the integers.

"Linked list" traverse the integers as follows:

    //array is full of 1-billion numbers, randomly sorted
    uint32_t idx = 0;
    for(int i=0; i
Pull out a stopwatch (or use Linux's "time" functionality). Divide the time by 200000000. You've now measured memory latency.

Re: AMD Ryzen Threadripper 3000 32-core CPU is more bad news for Intel

#143

Post by floatboth »

luizfelberti wrote: post #118
dragontamer wrote: post #25

Earlier quoted context omitted.

The main selling point for Intel isn't speed anymore. Intel still has superior performance counters and debugging features. Mozilla's rr (Record and Replay framework) only works on Intel for example, and Intel vTune is a very good tool. AVX512 is also an advantage, as you've noted. There are other instruction set advantages: I think Intel has faster division / modulus operator, and also has single-clock pext / pdep (…

Is AVX512 really an advantage though? Intel is infamous for severely downclocking the processor for these and other AVX/SSE family instructions, to a point where sometimes using them makes the program slower than it would be otherwise, especially if you're constantly provoking frequency switches between them and regular instructions. AMD might not have implemented AVX512 specifically yet (there's nothing legally keep…

> sometimes using them makes the program slower than it would be otherwise

oh no, that program will be super fast.

It would make other programs that run concurrently with the AVX one slower.

Re: AMD Ryzen Threadripper 3000 32-core CPU is more bad news for Intel

#145

Post by yummypaint »

ethanpil wrote: post #114

I'm ignorant in stock trading, I admit. But I dont understand why AMD stock isn't blowing up. It's done a recovery after a market wide dip, but hasn't all this Intel beating news raised the hype? So many other tech companies are so severly overvalued and hyped daily. Why is AMD a steady 28-32?

Considering how undervalued they have been for the last decade, things have moved definitively in the right direction. The recent releases and developments are still fresh news and i suspect the people/computers responsible for most of their volume operate at a longer timescale

Re: AMD Ryzen Threadripper 3000 32-core CPU is more bad news for Intel

#146

Post by yummypaint »

paulmd wrote: post #141
loeg wrote: post #122

Earlier quoted context omitted.

Wouldn't the higher clocks impact single-core performance? The article claims that's about the same. I suppose the userbenchmark single-core perf test[1] (the article links to some userbenchmark.com results) could just be a bad test; I'm unfamiliar with it. (It could also be that 2990WX wasn't power/heat limited on a single core, and neither is 3xxx; but say, at 16 core 100% clock, 3xxx can go higher than 2990WX due…

Zen2 is moving to a somewhat deceptive advertising model for clocks. AMD is advertising the absolute highest that any core on the die can hit under extremely light load for an absolute instant. Sustained single-core clocks will be 100-200 MHz less and sustained all-core will be significantly less. Most cores on a chip are not capable of sustaining the advertised clockrate even under severe voltage and even for instan…

Indeed, it makes it difficult to know what you're actually buying when considering something like a 3600 vs 3600x

Re: AMD Ryzen Threadripper 3000 32-core CPU is more bad news for Intel

#147

Post by NicoJuicy »

ethanpil wrote: post #114

I'm ignorant in stock trading, I admit. But I dont understand why AMD stock isn't blowing up. It's done a recovery after a market wide dip, but hasn't all this Intel beating news raised the hype? So many other tech companies are so severly overvalued and hyped daily. Why is AMD a steady 28-32?

It took a while to get the jump from 10 to 30. I actually was unlucky for selling just before, as it took too long ( holding it for a year).

I see it going to it's old peak soon. And it will just be as swift as it was before

Re: AMD Ryzen Threadripper 3000 32-core CPU is more bad news for Intel

#148

Post by archi42 »

The article references a Geekbench score that's now taken offline.

Google Cache: http://webcache.googleusercontent.com/search?hl=de&ei=ewaGXa...

Results (also in the article): Geekbench v5.0.1 Tryout for Windows x86 (64-bit); 1275 single core, 23015 multi-core for 32C/64T @ 3.59GHz Base Frequency (and with 32GB DDR4, no info on 4 or 8 channel).

Re: AMD Ryzen Threadripper 3000 32-core CPU is more bad news for Intel

#149

Post by PixelOfDeath »

nwmcsween wrote: post #60
dragontamer wrote: post #25

Earlier quoted context omitted.

The main selling point for Intel isn't speed anymore. Intel still has superior performance counters and debugging features. Mozilla's rr (Record and Replay framework) only works on Intel for example, and Intel vTune is a very good tool. AVX512 is also an advantage, as you've noted. There are other instruction set advantages: I think Intel has faster division / modulus operator, and also has single-clock pext / pdep (…

It also has superior vulnerabilities

NSA recommends Intel!

Re: AMD Ryzen Threadripper 3000 32-core CPU is more bad news for Intel

#150

Post by elheffe80 »

dragontamer wrote: post #83
elheffe80 wrote: post #64

But what processor is best for Dwarf Fortress?

Dwarf Fortress is IIRC mainly limited by RAM latency. So you really should be trying to get 3600 MHz RAM with as tight timings (maybe CL16 timings?) as possible. Dwarf Fortress's simulation is all about pointer-indirection and jumping around memory. The CPU doesn't really do much except wait for RAM most of the time. It takes ~50ns to talk to RAM, but the CPU is clocked at 4GHz (0.25 nanoseconds), giving you an idea…

Awesome! Thanks for the write up and good info.
Post reply on HN