• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

*** AMD "Zen" thread (inc AM4/APU discussion) ***

Choo Choo hype train!
6c12th Ryzen CPU at stock (up to 3.8Ghz), showing around 10-15% ST performance improvement versus 2500K running at 4.4Ghz, MT performance improvement around 100%

Will be really interesting to see performance on these things once clocked, especially for those of us with older CPUs.

Think my [email protected] is around 1800-1900 ST and 9K MT.
 
I'm thinking of getting one of the 8c16t chips to go in my Plex Server instead of my original plan to upgrade my 4670k to a 4790k. Mainly owing to the fact the i7 is massively overpriced second hand. I just hope this is not a big disappointment, it's about time the CPU market had a big shake up.
 
Alexrose1uk said:
OK if that's true, then this CPU is looking pretty epic based on this older bench run I've seen of a 5820K (Source: Techpowerup https://www.techpowerup.com/forums/threads/share-your-cpuz-benchmarks.216765/):

That Ryzen being CPU-z'd is also 6c12t and much lower clocked than the 5820k!
Click to expand...

The only thing I might say is, try and get the same version of cpu-z incase there is any difference from one version to another. But the scores look good, in terms of single thread vs multithread score, that Zen result looks like it scales incredibly. Where your 5820k and TaKen's 2500k both get less than single thread x core count, the Zen gets over single thread x core count. Both the 5820k and Zen have HT so it should be 12 threads a piece but the Zen is certainly getting better scaling.


From looking at a couple cpu-z results, a 6700k and therefore 7700k, seems to get around 2500/10000 score at basically 5ghz.


https://linustechtips.com/main/topic/430106-post-your-cpu-z-bench-scores/?page=5

If you scroll down there is a list of results. 4790k @ 4.85Ghz 2193/9525, 6700k @ 4Ghz 2059/8727, 6700@ 3.4Ghz 1830/7840, 6600 @ 3.3Ghz 1823/?.

So if the benchmark ran on that Zen at around 3.3Ghz, it has (in this benchmark) extremely competitive single thread and genuinely excellent MT scaling. Even if it's boosting to around 3.5Ghz on all cores, it's still extremely good performance.

That list of results shows better MT scaling on the 4970/6700's than the 5820k results above though.
 
drunkenmaster said:
So if the benchmark ran on that Zen at around 3.3Ghz, it has (in this benchmark) extremely competitive single thread and genuinely excellent MT scaling. Even if it's boosting to around 3.5Ghz on all cores, it's still extremely good performance.
Click to expand...

Has it been confirmed if the boost is on all cores or a single core?
 
Alexrose1uk said:
Choo Choo hype train!
6c12th Ryzen CPU at stock (up to 3.8Ghz), showing around 10-15% ST performance improvement versus 2500K running at 4.4Ghz, MT performance improvement around 100%

Will be really interesting to see performance on these things once clocked, especially for those of us with older CPUs.
Click to expand...

Very much so, seeing as that's about what my 2500k wants to deliver these days. Single-thread performance is as valuable to me as multi, so if I can get more cores AND a significant ST improvement, then that's exactly the sort of thing that would tempt me to finally upgrade :) If there's a reasonable chance of a 20-25% overclock, then the excitement shall happen.
 
Wasn't expecting a claim of better single threaded at lower clocks.

All the same, can't put too much weight on low detail numbers outside a proper review.
 
Odd Job said:
Has it been confirmed if the boost is on all cores or a single core?
Click to expand...

I don't think they'll give an all cores boost number, there is just too much variation. Intel doesn't do it either as it's just too complex to do.

The 6900k has a base clock speed of 3.2Ghz and a single thread boost of 3.7Ghz. All this is saying is in the worst case scenario, the most power draining application possible all cores will run 3.2Ghz, then the absolute top speed of a single thread is 3.7Ghz.... it's not making a statement at all about what the best turbo is for more than one thread, nor is it stating it will only run 3.2Ghz with 8+ threads. All cores can turbo to anything between 3.2-3.7Ghz and it depends on the application. Keep in mind from what I've seen the 6900k will hit 3.5Ghz on all cores for most people and it ran 3.5Ghz under boost in that handbrake comparison with Zen, despite having no official all cores boost.

But those are the two simple metrics, a single thread won't go above 3.7Ghz and all cores won't go below 3.2Ghz. I expect AMD will do the same, so for that 1800x likely 3.6Ghz base clock 4Ghz single core turbo, then some applications will run 8 cores at 3.8Ghz, some at 3.7Ghz, some at 3.6Ghz. Being an 8 core in a 95W TDP it's likely to do less well in terms of boost on all cores strictly within TDP. If you overclock, which might just be a case of raising the TDP you want to use, it might boost all the way to 4Ghz in everything.
 
trying to see which Vendor this chip was on, but then again would make it easier from AMD to track down who broke their NDA .
glad they did though, but least they could have finished installing all the drivers first before doing the shots haha
 
CPU-Z benchmark can be quite inconsistent between builds, but the scores look ok. I can only imagine the guys that are leaking these have a limited window to do so, or else they'd leak something that actually means anything. Like a valid CR15 score. Between this and PCMark, it makes matters even less transparent lol
 
My Sandy Bridge CPU:

bench2.png
bench.png
 
Tha
drunkenmaster said:
I don't think they'll give an all cores boost number, there is just too much variation. Intel doesn't do it either as it's just too complex to do.

The 6900k has a base clock speed of 3.2Ghz and a single thread boost of 3.7Ghz. All this is saying is in the worst case scenario, the most power draining application possible all cores will run 3.2Ghz, then the absolute top speed of a single thread is 3.7Ghz.... it's not making a statement at all about what the best turbo is for more than one thread, nor is it stating it will only run 3.2Ghz with 8+ threads. All cores can turbo to anything between 3.2-3.7Ghz and it depends on the application. Keep in mind from what I've seen the 6900k will hit 3.5Ghz on all cores for most people and it ran 3.5Ghz under boost in that handbrake comparison with Zen, despite having no official all cores boost.

But those are the two simple metrics, a single thread won't go above 3.7Ghz and all cores won't go below 3.2Ghz. I expect AMD will do the same, so for that 1800x likely 3.6Ghz base clock 4Ghz single core turbo, then some applications will run 8 cores at 3.8Ghz, some at 3.7Ghz, some at 3.6Ghz. Being an 8 core in a 95W TDP it's likely to do less well in terms of boost on all cores strictly within TDP. If you overclock, which might just be a case of raising the TDP you want to use, it might boost all the way to 4Ghz in everything.
Click to expand...

Thanks for that. I have been away from this for 10+ years and tech has moved on so much. Lots to learn again.
 
Far be it for me to throw a damp rag on everyone's hopes for this chip but has there been any more confirmation or updates/leaks on the Ryzen's rather high memory latency? From what I've seen so far that's Ryzen's only apparent weakness although I'm not sure how much of an impact that will have in real world applications.
 
You must log in or register to reply here.
Back
Top Bottom