• 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) ***

humbug said:
Thats a pretty contentious statement :p "the A64 destroyed the P4 but did not help AMD as none of the OEM's would use them."

Intel paid OEM's not to use AMD, thats an established fact, Intel had the book thrown at them for it to the tune of $2BN ordered to pay AMD.
Click to expand...

I am aware of the what happened but Intel made far more than it paid out and they are still fighting the case. From what I remember AMD got 1B and the EU collected 1B as a fine. I really hope AMD can get passed the Intel only mentality that a lot of companies/users have.
 
AMD Zen Naples Dual CPUs server ES with 64C/128T benchmark revealed.

http://wccftech.com/amd-zen-naples-soc-benchmarks/

Found Zen Naples 64C/128T benchmark tests result below:

http://browser.primatelabs.com/v4/cpu/105227

Zen Naples ES clocked 1.44GHz single core thread scored 1141 and multi core thread scored 15620.

If Zen Naples used 1 CPU then multi core thread would scored about 7810.

I found list of CPUs benchmarks score to compare below:

http://browser.primatelabs.com/processor-benchmarks

Zen single CPU core thread performance sound not impressed that it is slight faster than Atom Cherry Trail Z8700 when clock to clock. Zen Naples 32 cores multi thread at 1.44GHz would matched peformance as 3.3GHz 8 core AMD FX-8300 scored 7955. A 45W TDP Broadwell 16 core Xeon D 1571 CPU at 1.3GHz would eat 180W TDP 32 core Zen Naples for Geekbench lunch.
 
Yeah I saw that yesterday. Doesn't add up at all with AMD's claims and demo. Either it's nonsense or something very odd is going on with that benchmark.
 
I don't think you can read too much into it yet. The AMD geekbench entry only lists 64 cores (not 64 cores, 128 threads) like all other results, and isn't showing any L3 cache.

Within realms if just being early microcode / firmware/ drivers etc. Certainly not worth reading too much into yet.
 
DragonQ said:
Not sure why anyone is excited about Bristol Ridge, it has nothing to do with Zen at all.
Click to expand...

Well I can't argue why it should be appearing in the Zen thread other than that it shares a socket so its release has bearing on Zen news, but as for excitement, it is a chip that can meet normal people's usage all-in-one at fairly low power and will be very competitively priced. Assuming Intel don't start paying people not to use AMD chips again, this should be quite a coup for AMD.
 
Armageus said:
I don't think you can read too much into it yet. The AMD geekbench entry only lists 64 cores (not 64 cores, 128 threads) like all other results, and isn't showing any L3 cache.

Within realms if just being early microcode / firmware/ drivers etc. Certainly not worth reading too much into yet.
Click to expand...

Ahhh I see, didn't realised HT was disabled.

I ran 2 benchmarks yesterday with both HT enabled and disabled to see if there are difference in performance.

HT Disabled

https://browser.geekbench.com/v4/cpu/162978

HT Enabled

https://browser.geekbench.com/v4/cpu/161190

With HT enabled added about 1100 to multi core score for 8 threads, I figured out how Zen Naples would performed if HT was enabled.

Dual Zen Naples has 128 threads so 1100 x 16 = 17600 + 15620 = 33220

If HT was enabled Dual Zen Naples would scored about 33220 in multi core test. I thought it was impressed score but I found benchmark with dual 32 core 64 threads Intel Xeon E5-2683 v4 clocked 1.63GHz which is closer to Zen Naples 1.44GHz.

http://browser.primatelabs.com/v4/cpu/109880

I googled Intel Xeon E5-2683 v4, it is based on Broadwell EP, each 16 core 32 thread CPU used 120W TPD would absolutely destroyed Zen Naples in both single and multi core tests.

Made me wondered if AMD cherry picked Blender the only benchmark AMD can find and ran custom workload that saw Zen outperformed Broadwell-E clock to clock by 2% had everybody fooled.

If that how the final Zen will performed then AMD let everybody down again!
 
AthlonXP1800 said:
Ahhh I see, didn't realised HT was disabled.

I ran 2 benchmarks yesterday with both HT enabled and disabled to see if there are difference in performance.

HT Disabled

https://browser.geekbench.com/v4/cpu/162978

HT Enabled

https://browser.geekbench.com/v4/cpu/161190

With HT enabled added about 1100 to multi core score for 8 threads, I figured out how Zen Naples would performed if HT was enabled.

Dual Zen Naples has 128 threads so 1100 x 16 = 17600 + 15620 = 33220

If HT was enabled Dual Zen Naples would scored about 33220 in multi core test. I thought it was impressed score but I found benchmark with dual 32 core 64 threads Intel Xeon E5-2683 v4 clocked 1.63GHz which is closer to Zen Naples 1.44GHz.

http://browser.primatelabs.com/v4/cpu/109880

I googled Intel Xeon E5-2683 v4, it is based on Broadwell EP, each 16 core 32 thread CPU used 120W TPD would absolutely destroyed Zen Naples in both single and multi core tests.

Made me wondered if AMD cherry picked Blender the only benchmark AMD can find and ran custom workload that saw Zen outperformed Broadwell-E clock to clock by 2% had everybody fooled.

If that how the final Zen will performed then AMD let everybody down again!
Click to expand...

Why?
 
AthlonXP1800 said:
Ahhh I see, didn't realised HT was disabled.

I ran 2 benchmarks yesterday with both HT enabled and disabled to see if there are difference in performance.

HT Disabled

https://browser.geekbench.com/v4/cpu/162978

HT Enabled

https://browser.geekbench.com/v4/cpu/161190

With HT enabled added about 1100 to multi core score for 8 threads, I figured out how Zen Naples would performed if HT was enabled.

Dual Zen Naples has 128 threads so 1100 x 16 = 17600 + 15620 = 33220

If HT was enabled Dual Zen Naples would scored about 33220 in multi core test. I thought it was impressed score but I found benchmark with dual 32 core 64 threads Intel Xeon E5-2683 v4 clocked 1.63GHz which is closer to Zen Naples 1.44GHz.

http://browser.primatelabs.com/v4/cpu/109880

I googled Intel Xeon E5-2683 v4, it is based on Broadwell EP, each 16 core 32 thread CPU used 120W TPD would absolutely destroyed Zen Naples in both single and multi core tests.

Made me wondered if AMD cherry picked Blender the only benchmark AMD can find and ran custom workload that saw Zen outperformed Broadwell-E clock to clock by 2% had everybody fooled.

If that how the final Zen will performed then AMD let everybody down again!
Click to expand...

Blender is usually very Intel friendly.

There is too much speculation, the architecture is still in its testing phase, we don't know what speed it was running at, how many cores / threads running, GeekBench is reading 64 cores, it reads an FX8300 as 8 cores when technically its a 4 core with hyper-threading so who knows how many cores and threads were actually running, all Geekbench knows is that its an AMD architecture, if it thinks it an FX CPU its reading 32 cores and 64 threads as 64 cores....

Wccf are assuming its the dual 32 core platform AMD showcased, its could just as easily be a single CPU platform.

We just don't know anything.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom