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

[CentroX]

Could some one with a 6700k or 7700k run the cpuz bench and post a picture of it please so I can see where it stands vs. what has been posted? Pretty pretty please?? Thanks!!

6700K at 4GHZ

 

[Phynaz]

Wouldn't be a problem if AMD simply internally gimped memory performance... Would be pretty easy to do, but they still wouldn't be able to hide their full performance.

How would it not be a problem? Now a MB maker has 250,000 boards that haven't been stability tested with the CPU that will be put in them by customers. What company is going to take that chance?

Edit: I think this would also cause the MB to have to resubmitted to the FCC for EMI certification.

 

[JoeRambo]

Could some one with a 6700k or 7700k run the cpuz bench and post a picture of it please so I can see where it stands vs. what has been posted? Pretty pretty please?? Thanks!!

5Ghz 7700K is getting 2500 / 10641

 

[Justinbaileyman]

6700K at 4GHZ

Thank you very much!! So Skylake and Kabylake wins on single thread performance ever so slightly vs. Ryzen but looses horribly on multi thread performance.. I had just purchased a i7 7700k over the weekend too, now thinking of returning since I haven't opened it yet and instead going with 8c/16t Ryzen R1800x.. I wish there were more Ryzen benchies I could go on so I could make up my mind what to do.

 

[richierich1212]

Thank you very much!! So Skylake and Kabylake wins on single thread performance ever so slightly vs. Ryzen but looses horribly on multi thread performance.. I had just purchased a i7 7700k over the weekend too, now thinking of returning since I haven't opened it yet and instead going with 8c/16t Ryzen R1800x.. I wish there were more Ryzen benchies I could go on so I could make up my mind what to do.

Should've waited for official reviews of Ryzen since it was going to launch soon. It all depends on if you need to build your system now.

If not, there's no harm in waiting for Ryzen and/or Intel's response.

 

[Shivansps]

That CPU-Z bench is really impressive, especially if that's a 65w part.

Ryzen 6/12 3.3 - 3.7Ghz

ST Score: 1888
MT Score: 12544

My 5820K @ 4.25GHz

ST: 1929
MT Score: 11872

Amazing if true. I was planning on buying the cheapest 6/12 Ryzen CPU for my Plex server, but man; this is making me want to replace my Main rig with the 8/16 variant...

I'll run the test at stock speeds when I get the chance. Probably tonight

But thats a 1600 or a 1600X? be carefull with X and their XFR when comparing benchs, specially very short benchs like that one, it could be running way over 4ghz for all we know.

 

[imported_jjj]

But thats a 1600 or a 1600X? be carefull with X and their XFR when comparing benchs, specially very short benchs like that one, it could be running way over 4ghz for all we know.

It could be but not really if you factor in the multi threading result.
Even at equal clocks for turbo single vs all cores, the scaling is very good. If ST turbo would be higher, the MT result would be inexplainable.

Anyone got a clue how this bench scales with HT?

 

[Justinbaileyman]

I really really am wanting the 8c/16t 1800x 3.6-4.0ghz model now that I've seen what a 6c/12t can do. The only thing holding me back from returning my 7700k is what if when released the price actually ends up being $800 or $1000 instead of the rumored $450-$500.00. I am hoping to get a CPU/Mobo combo deal for $700-$750 tops since I already have DDR4 ram raring and waiting to go.

 

[Shivansps]

It could be but not really if you factor in multi threading result.
Even at equal clocks for turbo single vs all cores, the scaling is very good. If ST turbo would be higher, the MT result would be inexplainable.

XFR auto-overclocks it over the turbo clocks, we cant really be sure of the frecuency on ST or MT, it could be 4.2 ST, 4 MT, or more.

 

[imported_jjj]

I really really am wanting the 8c/16t 1800x 3.6-4.0ghz model now that I've seen what a 6c/12t can do. The only thing holding me back from returning my 7700k is what if when released the price actually ends up being $800 or $1000 instead of the rumored $450-$500.00. I am hoping to get a CPU/Mobo combo deal for $700-$750 tops since I already have DDR4 ram raring and waiting to go.

Some retailers will take pre-orders next week, confirming pricing ahead of retail availability on March 2.
Remains to be seen how fast it sells out , if it delivers on perf and price.

 

[bjt2]

But thats a 1600 or a 1600X? be carefull with X and their XFR when comparing benchs, specially very short benchs like that one, it could be running way over 4ghz for all we know.

I recall the CPUz screen and the string was an ES... AFAIK all ES around had XFR disabled...

 

[IEC]

XFR auto-overclocks it over the turbo clocks, we cant really be sure of the frecuency on ST or MT, it could be 4.2 ST, 4 MT, or more.

That's not correct per people who are under NDA and can't disclose more.
https://forums.anandtech.com/threads/speculation-on-ryzen-overclocking.2499538/#post-38744062

 

[imported_jjj]

XFR auto-overclocks it over the turbo clocks, we cant really be sure of the frecuency on ST or MT, it could be 4.2 ST, 4 MT, or more.

You've missed the point.

http://www.guru3d.com/articles_pages/core_i7_6950x_6900k_6850k_and_6800k_processor_review,9.htm


6700k scores 2118 and 9216.
So the 6700k scales well, assuming 4GHz for all cores and 4.2GHz for single, single core at 4GHz would be 2017 and 4 cores at 4GHz would be 8068 at perfect scaling. That implies that HT adds over 14%.
The 6950X at 4.4GHz does 2123 and 22222 so HT adds 4.7%

Ryzen scored 1888 and 12644 .
At same clocks ST and MT , perfect scaling would be 11328 and HT would add 12%.
If you assume 4GHz ST and 3.7GHz MT, it would mean that HT/SMT adds 21% over perfect scaling and isn't that a bit unrealistic?
Sure, the XFR all cores could be there even if we assume it isn't but a large diff between ST and MT clocks seems unlikely.

Would be nice to have some results with HT enabled and disabled for this bench.

 

[Greyguy1948]

Quotes:
It could be but not really if you factor in the multi threading result.
Even at equal clocks for turbo single vs all cores, the scaling is very good. If ST turbo would be higher, the MT result would be inexplainable.

Anyone got a clue how this bench scales with HT?

#5782imported_jjj, 25 minutes ago ....end of Quotes

HT is not working very good for Intel. You get some ref CPU if you run the test.
Core i5 is much better than Core i3
It would be very interesting to see 8C/8T, 6C/6T and 4C/4T for Ryzen
For games they could be the winners....

 

[Dygaza]

Anyone got a clue how this bench scales with HT?

Just with affinity:

I7-3770k @4.1GHz

4C/4T
Single 1774
Multi 6714

4C/8T
Single 1772
Multi 7641

 

[The Stilt]

The CPU-Z benchmark uses legacy code. The benchmark it runs is compiled with extremely old MSVC 2008, without SSE2 instructions being enabled.
In this type of a workload Zen is extremely strong.

 

[IEC]

I really really am wanting the 8c/16t 1800x 3.6-4.0ghz model now that I've seen what a 6c/12t can do. The only thing holding me back from returning my 7700k is what if when released the price actually ends up being $800 or $1000 instead of the rumored $450-$500.00. I am hoping to get a CPU/Mobo combo deal for $700-$750 tops since I already have DDR4 ram raring and waiting to go.

Fry's - $489 for 1800X per employee
http://imgur.com/HciQqW5

Could be fake I suppose, but fits with the rumored pricing / ShopBLT leaks.

 

[imported_jjj]

Just with affinity:

I7-3770k @4.1GHz

4C/4T
Single 1774
Multi 6714

4C/8T
Single 1772
Multi 7641

Thanks!

So HT adds some 14% for your IB.
The scaling without HT is at 0.946 on a per core basis and with HT at 1.078.
Ryzen scales at 1.116 with HT. If without HT it would scale same as yours, it would do 10716 at same ST and MT clocks so that implies that HT adds 18%.
If ST clocks would be higher than MT, AMD has some crazy good SMT.

 

[Shivansps]

That's not correct per people who are under NDA and can't disclose more.
https://forums.anandtech.com/threads/speculation-on-ryzen-overclocking.2499538/#post-38744062

Im sorry, but Extended Frecuency Range cant mean anything else than going over tubo clocks, what you understand for range and extended range? Or maybe they name it wrong.


BTW, it looks like it was an 1500 not a 1600X

 

[ComputerGuy2006]

Getting about same IPC but with two the core count = win. 4 cores is just not enough power for my daily task in 2017.

Even tho those numbers look really promising this entire situation reminds me of the hype AMD was trying to pull with bulldozer. Hope it works out and its not AMD playing games again.

 

[Justinbaileyman]

Me too.. I know all to well about the AMD hype wagon. They got me with the 939 x2 3600+,AM2 X4 940,AM3+ x4 fx8100,and Fm2+ x4 760k. Was not to happy with any of them. I did however really like the AM3+ FX-8300 and my Am1 5370 rigs. Both were cheapo buys and both overclock like champs..

 

[Janooo]

Just for reference:

at

 

[itsmydamnation]

The CPU-Z benchmark uses legacy code. The benchmark it runs is compiled with extremely old MSVC 2008, without SSE2 instructions being enabled.
In this type of a workload Zen is extremely strong.

Anything that isn't massively 256bit AVX1/2/FMA should all fit in that category.

 

[inf64]

Anything that isn't massively 256bit AVX1/2/FMA should all fit in that category.

I wonder what will happen on the reviews day, what kind of workloads will be used by certain websites .

 

[Justinbaileyman]

Yeah but that 6900k is 140w and tops out around 180w-200w when overclocked. I had a x99 set up this last summer and got rid of it asap due to major electric bill increases and soon as I switched to my current system which is skylake based my electric bill drop quite a bit wouldn't ya know. Thats why I want a Ryzen rig cause the TDP is only 95w..