ZEN ES Benchmark from french hardware Magazine

[The Stilt]

Why are you so defensive? it's pretty clear now that Zen's IPC will be above Sandy and Ivy's.

You are quoting old estimations which were based on the known performance of an existing architecture and an official statement from AMD (40% IPC over Excavator). When the estimation becomes inaccurate simply because the official statement from AMD was inaccurate, it is somehow the fault of the person (me) who used the available information?

 

[Maverick177]

And why are you being so offensive against him? Those predictions he made were based on data available back then and they were reasonable and in line with what AMD had said at the time.

Like I said, I have nothing against him. All I want is consistency.

You are quoting old estimations which were based on the known performance of an existing architecture and an official statement from AMD (40% IPC over Excavator). When the estimation becomes inaccurate simply because the official statement from AMD was inaccurate, it is somehow the fault of the person (me) who used the available information?

Ok here's the thing.

You 2 claimed Zen's IPC would be on par with Sandy and Ivy's IPC back in August based on the +40% IPC uplift claim from AMD. You also claimed your estimation is wrong is because AMD made inaccurate statements.

Don't you think AMD would claim " closer to 50% IPC gain " back then and now?

 

[The Stilt]

Like I said, I have nothing against him. All I want is consistency.



Ok here's the thing.

You 2 claimed Zen's IPC would be on par with Sandy and Ivy's IPC back in August based on the +40% IPC uplift claim from AMD. You also claimed your estimation is wrong is because AMD made inaccurate statements.

Don't you think AMD would claim " closer to 50% IPC gain " back then and now?

Sorry, but I don't understand what you mean.

If Zen delivers 40% more IPC than Excavator, then it matches Sandy / Ivy Bridge. That hasn't changed and won't ever change.
However if it infact matches Broadwell-E's IPC, then the actual increase over Excavator is ~55% and not 40% like AMD stated.

 

[Abwx]

I'll give you a hint: 40% IPC increase over Excavator matches almost exactly the IPC of Sandy / Ivy Bridge. Not Broadwell-E's, which appears to be the case.

.

Because you are stuck in FP based benches wich are a special case given the CMT uarch, what about looking at Integer based code.?.

FI Haswell has 40% IPC advantage compared to Piledriver in Fritzchess, and even less in more recent chess games, in 7ZIP it s well below 40%, so with 40% improvement over XV Zen IPC would be higher than HW in most INT based loads..

Also CanardPC explained their 35% better perf than Piledriver at equal frequency, the comparison is a Zen core with SMT and a PD module, so basicaly a Zen core throughput is 2.7x the one of a CMTed PD core, with 0.83x CMT penalty this still yield 2.24x the throughput of a PD core, if SMT gain is 30% then IPC improvement over PD is 1.72x...

35% à fréquence égale

La comparaison concerne un seul core avec CMT ou SMT dans les deux cas (c’est à une dire 1 core / 2 threads pour Zen et 1 core / 2 clusters pour le FX-8370)

http://www.cpchardware.com/cpc-hardware-n31-precisions-elucubrations/

And btw :

ZenOC@Air=5G

One core loaded and a big air cooler, MB VRMs werent beefy enough to try with all cores.

 

[Maverick177]

Sorry, but I don't understand what you mean.

If Zen delivers 40% more IPC than Excavator, then it matches Sandy / Ivy Bridge. That hasn't changed and won't ever change.
However if it infact matches Broadwell-E's IPC, then the actual increase over Excavator is ~55% and not 40% like AMD stated.

Ok so now I'm confused.

Do you still think Zen IPC's on par with Sandy and Ivy's ? Yes or No.

If No, then what's your estimation now?

 

[The Stilt]

Because you are stuck in FP based benches wich are a special case given the CMT uarch, what about looking at Integer based code.?

Of course I'm stuck with FP, simply because that's what has been lacking most on 15h.
Why concentrating to your strengths when you have morbid weaknesses? That would appear as self deception to me.

ps. Skylake is 107% ahead of XV in IPC in X265 and over 60% in VP9. They're both integer

 

[The Stilt]

Ok so now I'm confused.

Do you still think Zen IPC's on par with Sandy and Ivy's ? Yes or No.

If No, then what's your estimation now?

Of course not, if the performance demonstrated in the leaks and the results AMD has demonstrated are actually constant and true.
However in this case, the performance hasn't increased by 40% but around 55%.

AMD has officially only stated 40% and nothing else.

 

[CatMerc]

The hell are you guys trying to get out of The Stilt? He made an estimation using the data available to him, and the data ended up changing (AMD changed their claim from +40% to OVER 40%). Hence obviously that estimation won't end up being true if the data changed.

You're attacking him like rabid dogs. It's not like he's going full Juanrga and refusing to admit the possibility that he was wrong even when evidence goes against it.

 

[railroadmaster]

If it does overclock to 5ghz on air, I will be truly amazed. Everyone loves a winner and AMD will start winning again! The amazing comeback story like no other!

 

[.vodka]

AMD has officially only stated 40% and nothing else.

Lisa said at new horizon that they exceeded the 40% figure. Didn't state by how much. You could take that as another official statement. I mean, it's the CEO speaking. Leaks so far point at a healthy increase over the 40% figure.

You're attacking him like rabid dogs. It's not like he's going full *he who shall not be named* and refusing to admit the possibility that he was wrong even when evidence goes against it.

Oh hell no. Don't even mention that name, who knows, he may be summoned in some way or form.

 

[SlowSpyder]

I wouldn't be surprised if AMD used 40% as a nice round number. Remember, workloads and performance vary. Zen will probably be very strong in some areas, and closer to Sandy in others. Can't wait for launch, all signs are pointing to AMD having a winner, but we'll see.

 

[The Stilt]

Lisa said at new horizon that they exceeded the 40% figure.

She sure did, however it was rather hard to know that in advance
Until then the only figure had been 40%, as seen and heard in various presentations and slides.

 

[.vodka]

She sure did, however it was rather hard to know that in advance
Until then the only figure had been 40%, as seen and heard in various presentations and slides.

Oh, of course! We all had the 40% figure to work with until that moment, it was a shock to see the performance way above a projected 8C16T Sandy-E/Ivy-E. I completely agree with you.

One couldn't have predicted over 40% seeing the powerpoint slides with the high level uarch description with much certainty.

The overall picture has changed for the better.

 

[Dannotech]

Of course not, if the performance demonstrated in the leaks and the results AMD has demonstrated are actually constant and true.
However in this case, the performance hasn't increased by 40% but around 55%.

AMD has officially only stated 40% and nothing else.

At their New Horizon event, Lisa Su said they actually beat their original goal of 40% IPC uplift.

 

[crashtech]

Performance being better than promised is rather out of character for AMD lately, no wonder everyone is feeling blindsided, for some this is an unpleasant sensation, but it will make for an interesting 2017.

 

[zinfamous]

And why are you being so offensive against him? Those predictions he made were based on data available back then and they were reasonable and in line with what AMD had said at the time.

this.

people be crazy in this thread for absolutely no reason. On both sides.

 

[zinfamous]

Ok here's the thing.

You 2 claimed Zen's IPC would be on par with Sandy and Ivy's IPC back in August based on the +40% IPC uplift claim from AMD. You also claimed your estimation is wrong is because AMD made inaccurate statements.

Don't you think AMD would claim " closer to 50% IPC gain " back then and now?

Huh? Didn't Lisa say at NH: "We made our 40% target, 'and then some.'" Or something like that. She said that they went beyond their expectations.

 

[CatMerc]

Performance being better than promised is rather out of character for AMD lately, no wonder everyone is feeling blindsided, for some this is an unpleasant sensation, but it will make for an interesting 2017.

From what things sound like, AMD themselves got surprised by the gains. And according to Fottemberg with the latest revision clocks are higher than they expected too.
It's that Jim Keller and his wizardy I tell you!

 

[raghu78]

Right, I'm starting to think that CB score might be too optimistic and not reflective of overall performance, compared to Intel's quads at least. According to the latest benchmarks 8C/16T Ryzen at 3.2-3.3 GHz all core Turbo would edge stock 7700K with half the cores by only 10-12.5% in applications and still lose badly in CPU limited games - and that's before you overclock the latter to 5 GHz (easily done in many reviews). Go figure.

That's a lot of 'assumptions' considering that user has zero info about Ryzen prices, overclocking headroom and only a limited idea of how it performs. I'm particularly curious to see how the 6C/12T Ryzen 'will bury 7700K' prediction holds up.

Oh do you at all ? AMD Zen ES are not final production silicon. Anybody who benched using those including canard PC is not giving an accurate picture of performance. canard PC agreed that their sample had bugs with SMT and uop cache. I am guessing that AMD has final production silicon in their hands and are finalizing base clock and turbo boost speeds. One thing is sure you are pretty much going to be out of business by March 2017 as there would be no rumours to spread.oh btw do not worry Zen will do well in Cinbench too. Just wait for final reviews of retail chip.

 

[lolfail9001]

Oh do you at all ? AMD Zen ES are not final production silicon. Anybody who benched using those including canard PC is not giving an accurate picture of performance. canard PC agreed that their sample had bugs with SMT and uop cache. I am guessing that AMD has final production silicon in their hands and are finalizing base clock and turbo boost speeds. One thing is sure you are pretty much going to be out of business by March 2017 as there would be no rumours to spread.oh btw do not worry Zen will do well in Cinbench too. Just wait for final reviews of retail chip.

Man, did you not learn your lesson with Polaris? Being emotionally invested into hype hurts, and i am of opinion Zen will be a fine product.

 

[CatMerc]

https://twitter.com/BitsAndChipsEng/status/814894218544037892

 

[Abwx]

Huh? Didn't Lisa say at NH: "We made our 40% target, 'and then some.'" Or something like that. She said that they went beyond their expectations.

Once Zen uarch was disclosed we knew that a 4 ALUs design with a beefed FPU could hardly have lower IPC than SB/IB wich are both 3 ALUs designs, and that 40% wasnt the real figure to expect from such a design....

 

[PhonakV30]

according to French's game bench , Zen at 3.1~3.5 Ghz is below Core i5-6600 at 3.5~3.9 , [ 97.3 / 73.6 (FX-8350) = 32% ]
so 4cores/8 threads at 4.0~4.2Ghz will match 4790K or lower.So overall IPC is over 40%. [ 107 / 73.6 (FX-8350) = 47% ]

 

[raghu78]

Man, did you not learn your lesson with Polaris? Being emotionally invested into hype hurts, and i am of opinion Zen will be a fine product.

We have seen Zen being demoed with competitive performance and perf/watt against Broadwell-E . Whats happening with Zen is the opposite of what happened with Polaris. AMD made bombastic statements about Polaris only to underwhelm. With Zen AMD has been managing expectations well for the past 4 years and now when they are close to launch they are showing that they have exceeded their goals. I would say thats the exact opposite of what happened with Polaris. In fact I think the best is yet to come. I think AMD is reserving the best details for launch. Base and Max turbo clocks and max clocks. Anyway the signs are Zen is shaping up to be very competitive. We will see how it turns out at launch.

 

[zinfamous]

Man, did you not learn your lesson with Polaris? Being emotionally invested into hype hurts, and i am of opinion Zen will be a fine product.

what's the problem with Polaris? It is already better than its more expensive competitor, the 1060 6GB. Yes, there was TONS of hype before Polaris, but look at where it is now, and what people were saying about "where it would be" only 6 months ago.

I'm actually impressed.