New Ryzens and DAWBench.

[Kaine]

It seems a shame with all of these cores being available, and with a good clock speed that the maximum logical core issue doesn't seem to get much attention.

Most sequencers can handle 32 threads absolutely fine. A few can go higher, 1 or 2 notabley less although not always intentially.

This combined with the the maximum unique .dll issue

I'm aware of it, but I can't recall the last time anyone spoke to me about hitting it. Noel's comments in http://forum.cakewalk.com/VST-Limit-Rep ... 68760.aspx thread outline it as being linked to C++ libary linking.

I'm not so sure that a x299 successor will show up anytime soon. I've been looking around, and NOBODY likes this platform except us who are just drooling, it's far from critical success for Intel. if I were them I'd consider killing the HEDT line and just have i7's and xeons.

Errm, i7 X299 is HEDT.

i9's are just i7 HEDT's with more cores sandwiched in and I'm not sure Intel care in the slightest for a critical success, they are still shifting them by the bucket load

If anything it's always been a common pastime for tech guys to rag on the Intel "X" series for being the wrong side of the price/performance cliff face, the i9's are simply that with a shiney new badge.

[standalone]

If I'm not missing something there is something weird with the i5-8400 result in the Vi test. It's the only CPU that doesn't experience a big boost when going from 128 samples to 256 samples:

i5-8400, from 550 to 580

i5-8600k, from 680 to 920

I would expect that both i5 processors evolve in parallel.

[Kaine]

The test is looking for the total amount of processing that can be done within a given ASIO cycle, so it's possible the weakest (slowest) core just brought the rest down quicker than your expecting on the 8400, the might even be some headroom on the other cores still it can't use because of it. If the test hardware setup itself was skewed then the DB4 test would be looking odd as well, but that looks to have scaled fine, so I suspect it's something like the above and how the sequencer or Kontakt has load balanced with the available resources.

I'll add them to the retest list, but whilst it might look a little odd I'm not shocked by it.

[Kaine]

*double post - board hang during posting*

[Kaine]

*triple post - board hang during posting*

[standalone]

The test is looking for the total amount of processing that can be done within a given ASIO cycle, so it's possible the weakest (slowest) core just brought the rest down quicker than your expecting on the 8400, the might even be some headroom on the other cores still it can't use because of it. If the test hardware setup itself was skewed then the DB4 test would be looking odd as well, but that looks to have scaled fine, so I suspect it's something like the above and how the sequencer or Kontakt has load balanced with the available resources.

I'll add them to the retest list, but whilst it might look a little odd I'm not shocked by it.

Thanks for the response. I was hoping for a typo, something like 850 instead of 580

[acYm]

I'm curious about something, now. in the 7900x review you said "with the 7900X the internal chip latency has been superb. Even running at a 64 buffer we’ve been seeing 100% CPU load prior to the audio breaking up in playback, making this one of the most efficient chips I think I’ve possibly had on the desk."

obviously this is what we want, and why I've been looking into x299, but, what's the 8700K's behavior in that regard? I see that it's barely below the 7820x for instances and polycount, and so possibly a real good value. if it's too penalized by the lack of mesh in real-time low latency use though, that's a minus.

(I figured out there's nothing scary about running w7 on z370 so that's a new door opened up for me)

[Kaine]

It's fine, I normally only tend to comment if it drops below 95% with somewhere between 95% and 98% where i'd traditionally expect break up to start occurring. From my notes, it was decent at around 97%.

[acYm]

that's still quite outstanding I think (from the perspective of a dinosandybridgeosaurus user). alright, thanks.

good lord, that 5.3 chip is tempting!

I think I'm going for it

[Freaqpeak]

So... Any DAW tests available yet? ..

[poshook]

When you are serious in music composing and use a lot of mega samples libraries, AMD is a piece of crap. More than one CPU die ends in very poor memory-allocation performance. Intel's one die tech is absolutely superior to AMD in this field. New Ryzen 3000 will be the same bullshit tech - many dies and slow memory access. This affect also DSP performance at very low latencies.

[tooneba]

I'm going to click buy button if dawbench shows promising scores. Or I can go 9900K if it still outperforms them.

[Freaqpeak]

When you are serious in music composing and use a lot of mega samples libraries, AMD is a piece of crap. More than one CPU die ends in very poor memory-allocation performance. Intel's one die tech is absolutely superior to AMD in this field. New Ryzen 3000 will be the same bullshit tech - many dies and slow memory access. This affect also the performance at very low latencies.

Haha, you can't say that BEFORE seeing the bechmarks can you.
I have friends running AMD Threadrippers as their main DAWs and they totally happy with their systems. Although the weren't cheap. Now, those are previous models. Ryzen 3000 and Zen 2 is another.
Music composing doesnt equal using pre-made sample libraries you now. Some of us make our own sounds in real time....

[Liero]

When you are serious in music composing and use a lot of mega samples libraries, AMD is a piece of crap. More than one CPU die ends in very poor memory-allocation performance. Intel's one die tech is absolutely superior to AMD in this field. New Ryzen 3000 will be the same bullshit tech - many dies and slow memory access. This affect also the performance at very low latencies.

Haha, you can't say that BEFORE seeing the bechmarks can you.
I have friends running AMD Threadrippers as their main DAWs and they totally happy with their systems. Although the weren't cheap. Now, those are previous models. Ryzen 3000 and Zen 2 is another.
Music composing doesnt equal using pre-made sample libraries you now. Some of us make our own sounds in real time....

Poster has obnoxious way of saying that AMD performs somewhat worse than Intel
___in the very specific case of Kontakt used with big sample libraries_____, which is true. If you don't use a lot of Kontakt, Ryzen 3rd right now is definitely a much better option.

[poshook]

When you are serious in music composing and use a lot of mega samples libraries, AMD is a piece of crap. More than one CPU die ends in very poor memory-allocation performance. Intel's one die tech is absolutely superior to AMD in this field. New Ryzen 3000 will be the same bullshit tech - many dies and slow memory access. This affect also the performance at very low latencies.

Haha, you can't say that BEFORE seeing the bechmarks can you.
I have friends running AMD Threadrippers as their main DAWs and they totally happy with their systems. Although the weren't cheap. Now, those are previous models. Ryzen 3000 and Zen 2 is another.
Music composing doesnt equal using pre-made sample libraries you now. Some of us make our own sounds in real time....

Yes as I see the same issues. If you use PC for recording and mixing, there is no reason to avoid Ryzen's CPU. But even AMD uses new CCD, 8- core chiplets are still based on 2x 4-Core CCX with shared Infinity fabric. And 12-Core or 16-Core CPUs will have two CCDs, I expect the same issues as Threadrippers had