AMD Ryzen (Summit Ridge) Benchmarks Thread (use new thread)

[cytg111]

Based on my own experience, Windows 7 performs slightly better than Windows 10 (even on Ryzen).
Even with most of the bloatware disabled.

- No way.. Windows10 is slimmed down, optimized to the iron best performing windows kernel ever.. so i've been told by microsoft propaganda .

 

[inf64]

I don't think any ES had XFR enabled, let alone Turbo? So trying to call those ES's X models is laughable... I can't recall any ES's being discovered with those features enabled, but if someone wants to correct me that's fine.

I think Turbo and MACT was working as we can see from the results of R15 and CPUz. XFR on the other hand is going to be enabled only on select few models in the whole lineup.

 

[Agent-47]

We already know Passmark CPU score is very heavily influenced by Memory latency and bandwidth. I dont know why people are trying to take "IPC" means for this data without first normalizing for memory performance.

What people should be doing it trying to find comparable system configurations.

Only physics and PN is memory depended in the CPU mark suite

 

[Eknex]

Hype Hype Hype AMD

One question, how do I know if the Crosshair IV has better VRM than an ASRock X370 (cheaper), for example?
Are they really (as usual) better at more price, or are they practically the same?

I never see that information

 

[inf64]

Hype Hype Hype AMD

One question, how do I know if the Crosshair IV has better VRM than an ASRock X370 (cheaper), for example?
Are they really (as usual) better at more price, or are they practically the same?

I never see that information

I think Asus will have the best motherboards and that comes with a price. I don't see why would X370 board from MSI or Asrock be a bad choice even if it is not as good as Asus.

 

[OrangeKhrush]

Only physics and PN is memory depended in the CPU mark suite

That PN bench is so quad channel dependant.

 

[unseenmorbidity]

I think Asus will have the best motherboards and that comes with a price. I don't see why would X370 board from MSI or Asrock be a bad choice even if it is not as good as Asus.

Do they really have the best or so they have the best marketing, which in turn created their reputation as the best?

When your brand is perceived to be the best, then you can ask more for it. Even, if it's perceived superior quality isn't born out in reality.

I am not saying you are wrong. I am genuinely curious.

 

[cytg111]

Sorry I don't get what you're trying to say. All I know is that the 4790K base speed is 4GHz with 4.4GHz turbo while the 4770K is 3.5GHz base with 3.9GHz turbo.

was typing out an answer detailing what IPC means when I confused myself. A chips IPC can really only be determined at a specific speed and subsystem(board, ram speed etc.).. Right?
So the IPC of 4770 and 4790 is infact, different (unless subsystem scales accordingly).

 

[OrangeKhrush]

I have had so many issues with Asus, I am a MSI fan, that said I have tried out Asrock and they throw everything on

 

[inf64]

was typing out an answer detailing what IPC means when I confused myself. A chips IPC can really only be determined at a specific speed and subsystem(board, ram speed etc.).. Right?
So the IPC of 4770 and 4790 is infact, different (unless subsystem scales accordingly).

IPC is clock independent. Both chips have Haswell cores and baring some drastic DDR3 memory subsystem differences these two would perform the same if clocked the same. It is simple.

 

[lolfail9001]

IPC is clock independent. Both chips have Haswell cores and baring some drastic DDR3 memory subsystem differences these two would perform the same if clocked the same. It is simple.

Well, Passmark claims the opposite.

 

[GroundZero7]

Hype Hype Hype AMD

One question, how do I know if the Crosshair IV has better VRM than an ASRock X370 (cheaper), for example?
Are they really (as usual) better at more price, or are they practically the same?

I never see that information

Wouldn't worry about a VRM that has any heatsink on it. I'm running 160W on my Phenom II X6 on an Asrock 970 budget board with aluminum roofing nails epoxied on em

 

[inf64]

Well, Passmark claims the opposite.

What does passmark claim?

 

[bjt2]

So the 1800X CB R15 MT

1136 x 8/6 * 3.4/3.6 = 1431.......that SMT

Shouldn't it be 1136 * 8 / 6 * 3.6 / 3.4, since the clock of the 8c is 3.6 and the clock of 6c is 3.4?

 

[inf64]

Shouldn't it be 1136 * 8 / 6 * 3.6 / 3.4, since the clock of the 8c is 3.6 and the clock of 6c is 3.4?

Yep we covered that above.

 

[lolfail9001]

What does passmark claim?

That 4790k displays like 5-7% higher IPC than 4770k in single threaded test. It has fairly ridiculous margins of variance.

It also has higher IPC than 6700k. Out of thin air.

I believe you have cited Passmark dev on this one, but he did not give a proper explanation imho.

Anyways, now that we are under onslaught of information that Zen is good/great/excellent/whatever, it is time to think of a way AMD can mess this up.

 

[Doom2pro]

Wouldn't worry about a VRM that has any heatsink on it. I'm running 160W on my Phenom II X6 on an Asrock 970 budget board with aluminum roofing nails epoxied on em

Anyone remember the MSI 970A G43/G46 boards? They had heatsinks and LOADS of FETs and the board was rated as being 125W but they used horribly cheap FETs and the board could only handle max 95W and lots of people smoked their boards CPU and all simply by putting in 8 core FX CPUs and overclocking them!

 

[sm625]

Then why the hell is Haswell like 60-70% faster than Ivy? What is this sorcery?

Haswell has effectively a double size instruction decode queue for a single thread, double the number of branch ports, an extra INT port, double the load/store units, and double the cache bandwidth. All good things to have for a single heavily branching thread.

 

[inf64]

That 4790k displays like 5-7% higher IPC than 4770k in single threaded test. It has fairly ridiculous margins of variance.

It also has higher IPC than 6700k. Out of thin air.

I believe you have cited Passmark dev on this one, but he did not give a proper explanation imho.

Anyways, now that we are under onslaught of information that Zen is good/great/excellent/whatever, it is time to think of a way AMD can mess this up.

I cannot see anything suspicious , it is bang on correct(within margin of error).
https://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i7-4770K+@+3.50GHz
Turbo Speed: 3.9 GHz
Single Thread Rating: 2255
If it ran at 4.4Ghz Turbo: 2255 x 4.4/3.9=2544

https://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i7-4790K+@+4.00GHz
Turbo Speed: 4.4 GHz
Single Thread Rating: 2527

Difference between 4770K and 4790K in ST Turbo clock : 12.8%
Difference between 4770K and 4790K in Single Thread Rating: 12%

That is 1.128/1.12=1.006 or 0% margin of error. Passmark shows zero IPC difference between 4770K and 4790K, facts are facts.

 

[sm625]

What's up with that SSE score? Looks like AMD flipped Intel the bird!

It's just those particular baselines. Other ones look much better. I would post some examples but the Nazis flogged me the last time I posted passmark screenshots.

 

[ozzy702]

Anyone remember the MSI 970A G43/G46 boards? They had heatsinks and LOADS of FETs and the board was rated as being 125W but they used horribly cheap FETs and the board could only handle max 95W and lots of people smoked their boards CPU and all simply by putting in 8 core FX CPUs and overclocking them!

Yup. I blew up two of those piles... I really hope that the AM4 boards are all solid.

 

[inf64]

It's just those particular baselines. Other ones look much better. I would post some examples but the Nazis flogged me the last time I posted passmark screenshots.

Take note of (PT8) scores when you do the baseline search from passmark application. Older results are not directly comparable with PT9.

 

[Dresdenboy]

Looks like the same source that Dresdenboy and I were speculating on a month ago that everyone denounced as "fake". I tried to say it may not be at Overclock.net and they were still locking threads and editing posts.

I wanted to post the CB screenshot, but it's still in the OP.

I have little clue (because i did not touch it yet) but it is the only viable explanation of jump between Haswell and Ivy, don't you think? Well, either that or

http://www.realworldtech.com/haswell-cpu/4/

I just looked a bit around on the web. Indeed it seems to be related to software rendering stuff. Some not so new forum postings I found hinted at low gains. This might have changed by now, but there is a chance, that other uarch improvements might have helped as well. For example branch prediction, prefetchers, different queue sizes, etc.This is supported by this statement from one of the AT reviews:

This benchmark runs a Wii program that raytraces a complex 3D scene inside the Dolphin Wii emulator. Performance on this benchmark is a good proxy of the speed of Dolphin CPU emulation, which is an intensive single core task using most aspects of a CPU.

.

 

[lolfail9001]

Never mind.

That screen[ in OP] still makes little sense tbh.

 

[inf64]

Got this link from reddit "stalker" who lurked the baido forum 24/7:
http://imgur.com/a/Us02h

Also this image from baidu topic lists 4Ghz Zen as scoring 165pts in R15 :

Thus 3.7Ghz should be getting 152 (4% higher than what the 6C ES scored).

More:

Translated(rough):
- IPC is at the least Ivy Bridge-E and higher
- SMT for Ryzen is more efficient then Intel's HyperThreading
- Ryzen has no cold Bug
- Cinebench R15 hits 145 single thread @ 3.4GHz on ES, earlier models hit 130-140, retails should hit 140-150. Ryzen @ 3.4Ghz 8C/16T should score ~1500pts in R15
- Most Ryzen ES samples hit 4.3-4.5GHz MAX on Air with all core enabled
- Intel is testing out Skylake-X, and beats out current 6950X with 8C/16T because it can hit higher clocks.

^^Above translation taken from overclock.net forum topic(I added only small bit about MT score in R15, it was missing).