(Rumour) - Zen chips have already been tested, met all expectation, no significant bottlenecks

[AresKrieger]

Am I one of the only people here who doesn't care either way about this news? It's basically a report saying operation is normal, which means nothing has changed since our last zen update except the date. On top of which the source is questionable as it is a rumor, lol.

[patrickjp93]

in theory, if all AMDs claims are true. it CAN be better then skylake in single core performance.

 

in reality, i would say it will be between Ivy Bridge and Haswell.

 

Zen+ should be around skylake

No it can't. The math's been done. Without 15% higher clocks Zen won't break even with Skylake.

People are acting like Intel can really release a CPU twice as fast as Zen as soon as it comes out. Do you people not understand that the law of diminishing returns is already hitting Intel hard? We are reaching the point now where the gains we are getting each generation is starting to slow down. I have no doubt Intel has something up their sleeve to answer Zen, but i do not think it will be a magnitude of "2x faster".

 

Zen does not have to beat Skylake. Skylake is only 11% faster AT BEST over Haswell, and that's not in regard to gaming. DDR4 has improved Skylake's performance in area's such as compression. In gaming the number is closer to 5%. If AMD gets within spitting distance of Haswell with Zen, then they have a winner. Need i remind you people that hypertransport is fast enough to handle 4-way SLI without needing a PLX8747 bridge? This means you do not need a $400-$500 motherboard to do 4 way SLI. Quite an advantage, don't you think?

 

We also do not know the exact configuration of the FPU's, or if they can be merged into larger sets to handle AVX3. If they can, it will be yet another advantage that Zen would have. However, with more information being released, i feel less hopeful for this happening.

 

If we assume AMD isn't lying, and that Keller is as magical as people make him out to be, then that 40% number over excavator might just land AMD near Haswell. That is all it will take for the platform to be successful. That, and competitive pricing. If they release a 4 core, 8 thread Zen chip at i5 prices, that can handle 4-way SLI on cheap enough, then how can that possibly go wrong?

Hot chips, terrible stock coolers, buggy instructions, buggy chipset, poorly constructed motherboards, lacking AVX 512 for enterprise (they'll be laughed out of the room), releasing in the off-season when most server and laptop purchases have been made until the Christmas season, Kaby Lake reaching 4.6GHz for quads and 3.8 for octals as has been rumored, Kaby Lake bringing HMC as L4 cache for graphics as has also been rumored... A lot can go very wrong for AMD, and that's before we discuss GloFo's yield issues.

[Curufinwe_wins]

Lol well if the expectations are what they were originally (on par or worse than haswell) then amd still will have a ways to go.

[TechGod]

Lol well if the expectations are what they were originally (on par or worse than haswell) then amd still will have a ways to go.

Skylake isn't even that much better than Haswell...

[Curufinwe_wins]

Skylake isn't even that much better than Haswell...

Indeed. But 40% over bulldozer is at BEST almost to haswell, and AMD still has to overcome the stigma they have developed.

[patrickjp93]

Skylake isn't even that much better than Haswell...

Anywhere from 5% to 15% better depending on the benchmarks. Scientific benches are in Skylake's favor. Gaming is a break even point, not entirely surprising given Skylake has the lower single thread clock rate.

[TechGod]

Indeed. But 40% over bulldozer is at BEST almost to haswell, and AMD still has to overcome the stigma they have developed.

They said the 40% boost was over Excavator, not Bulldozer. 

Furthermore they also said the 40% boost was independent of the process node used. Going from 28nm to 14nm is gonna massively improve performance if what they said is true. 

[Curufinwe_wins]

They said the 40% boost was over Excavator, not Bulldozer.

Furthermore they also said the 40% boost was independent of the process node used. Going from 28nm to 14nm is gonna massively improve performance if what they said is true.

They literally said 40% ipc improvement. They never specified node.

I hope it is better. I want amd to compete. I'm just saying their expectations are low enough where it doesn't mean anything to mean they meet expectations (because the expectations are not competitive).

[patrickjp93]

They said the 40% boost was over Excavator, not Bulldozer.

Furthermore they also said the 40% boost was independent of the process node used. Going from 28nm to 14nm is gonna massively improve performance if what they said is true.

IPC increases are not dependent on node shrinks. I don't understand how anyone ever started believing they were...

[TechGod]

 

IPC increases are not dependent on node shrinks. I don't understand how anyone ever started believing they were...

Don't blame me. Some people started going on about how IPC is linked directly to the process node used and that AMD was lying with their claim. *Shrugs*

 

 

They literally said 40% ipc improvement. They never specified node.

 

I hope it is better. I want amd to compete. I'm just saying their expectations are low enough where it doesn't mean anything to mean they meet expectations (because the expectations are not competitive).

 

http://www.techpowerup.com/212315/amd-zen-offers-a-40-ipc-increase-over-excavator.html

 

Its 14nm apparently.

[Curufinwe_wins]

http://www.techpowerup.com/212315/amd-zen-offers-a-40-ipc-increase-over-excavator.html

Its 14nm apparently.

It's the tsmc 14nm (which Samsung calls 16nm), which is a notable amount larger node that intel's 14nm (Intel has really really tight nodes).

It will be small enough though. I meant that the ipc specification they noted was not with the indication that it was a specific node. (As in that somehow you might get an 40% architecture increase then an additional increase from node scaling.)

[TheProfosist]

Weren't there similar rumors for Bulldozer? And we all know how that train wreck ended.

@Arcanekitten The quote in your signature is literally invisible to Night Theme users, might want to remove the coloring on that text.

Bulldozer was hyped through the roof. Also Server guys still like Bulldozer, why wouldn't they?

[Dabombinable]

Bulldozer was hyped through the roof. Also Server guys still like Bulldozer, why wouldn't they?

Well it is cheaper to by the FX series than the Opterons.

[Prysin]

Indeed. But 40% over bulldozer is at BEST almost to haswell, and AMD still has to overcome the stigma they have developed.

NOT bulldozer.

 

Excavator.

 

If it was 40% over bulldozer, it would barely catch up to Sandy Bridge

[TechGod]

NOT bulldozer.

 

Excavator.

 

If it was 40% over bulldozer, it would barely catch up to Sandy Bridge

So annoying when people bash something they have barely researched...

[Prysin]

Hot chips, terrible stock coolers, buggy instructions, buggy chipset, poorly constructed motherboards, lacking AVX 512 for enterprise (they'll be laughed out of the room), releasing in the off-season when most server and laptop purchases have been made until the Christmas season, Kaby Lake reaching 4.6GHz for quads and 3.8 for octals as has been rumored, Kaby Lake bringing HMC as L4 cache for graphics as has also been rumored... A lot can go very wrong for AMD, and that's before we discuss GloFo's yield issues.

the only company with shit stock coolers are intel. Do not even try to argue this.

 

Intel cooler is shit. The end.

[AlwaysFSX]

Server people are happy? Great, we're getting shared resources again. Fucking wonderful if that turns out to be true.

 

the only company with shit stock coolers are intel. Do not even try to argue this.

 

Intel cooler is shit. The end.

Really only if you're being dumb and use it for an i7 or OC'd i5.

[Dabombinable]

the only company with shit stock coolers are intel. Do not even try to argue this.

 

Intel cooler is shit. The end.

No, they aren't look at the Pentium 4 stock cooler, then imagine it on a Pentium E6500K with its far lower TDP and greater performance. As for the current cooler, it works just fine on the locked CPU. BTW, I have the AM3+ cooler that's from my friends FX 8350 in one of my parts draws, so I can quite honestly compare them and say that AMD's coolers reflect the TDP of the CPU that they are being used on.

[Curufinwe_wins]

NOT bulldozer.

 

Excavator.

 

If it was 40% over bulldozer, it would barely catch up to Sandy Bridge

Indeed, my statement was wrong, the conclusion gained from it was not.

 

So annoying when people bash something they have barely researched...

Note that I did not propagate that statement whatsoever. In fact the only comments I made on it afterwards had nothing to do with that statement. You yourself also made completely wrong statements throughout our discussion, but I did not call you out for "bashing something you have barely researched" when clearly you (albeit in a different way) also misremembered (for I do not believe it was willful ignorance) their original statements.

 

I trust that you like I am at least willing to accept that mistakes are made and move forward. My offer on the previous discussion stands of course, and I am indeed glad to see it was added to your signature. Now the waiting game continues.

[TechGod]

Indeed, my statement was wrong, the conclusion gained from it was not.

 

Note that I did not propagate that statement whatsoever. In fact the only comments I made on it afterwards had nothing to do with that statement. You yourself also made completely wrong statements throughout our discussion, but I did not call you out for "bashing something you have barely researched" when clearly you (albeit in a different way) also misremembered (for I do not believe it was willful ignorance) their original statements.

 

I trust that you like I am at least willing to accept that mistakes are made and move forward. My offer on the previous discussion stands of course, and I am indeed glad to see it was added to your signature. Now the waiting game continues.

Yup. Will gladly admit if wrong. Sorry for getting aggresive. 

[Curufinwe_wins]

Yup. Will gladly admit if wrong. Sorry for getting aggresive. 

It's fine. Happens to all of us. @Prysin and I have had some pretty heated arguments in the past, although really most of them boil down to the more gray areas, and differences in perception rather than either of us being significantly off in our knowledge of reality.

 

 

the only company with shit stock coolers are intel. Do not even try to argue this.

 

Intel cooler is shit. The end.

An amusing video on the topic. 

 

Now in fairness to what has been mentioned already, it does reflect the lower expected TDP of the cpu's being used, but I will say straight out the intel one rides the acceptable limits much harder than the AMD cooler. Now if intel didn't do things like the original haswell tim every once in a while, I doubt I'd mind much, but any situation like that one is purely unacceptable.

 

 

[TheProfosist]

Well it is cheaper to by the FX series than the Opterons.

The architecture really. I mean good multithreading can only benefit servers. AMD had a higher core count as well for the longest time, actually they still may.

the only company with shit stock coolers are intel. Do not even try to argue this.

 

Intel cooler is shit. The end.

Slap a 1156 stock cooler on a 1151.

It's fine. Happens to all of us. @Prysin and I have had some pretty heated arguments in the past, although really most of them boil down to the more gray areas, and differences in perception rather than either of us being significantly off in our knowledge of reality.

 

 

 

An amusing video on the topic.

 

Now in fairness to what has been mentioned already, it does reflect the lower expected TDP of the cpu's being used, but I will say straight out the intel one rides the acceptable limits much harder than the AMD cooler. Now if intel didn't do things like the original haswell tim every once in a while, I doubt I'd mind much, but any situation like that one is purely unacceptable.

Just remember that AMD has lower limits than Intel does. Also generally their fans run at a higher RPM and are louder.

[Dabombinable]

The architecture really. I mean good multithreading can only benefit servers. AMD had a higher core count as well for the longest time, actually they still may.

Slap a 1156 stock cooler on a 1151.

I think the Opteron only goes to 8 modules, or 16 "cores".

[Curufinwe_wins]

Just remember that AMD has lower limits than Intel does. Also generally their fans run at a higher RPM and are louder.

AMD also uses fake temp sensors, supposedly reporting god only knows what temperature, where I have no fucking clue what their actual Tj max is. Since both are silicon and honestly material properties don't vary that much at the same lithography, I would expect Tj max of similar to Westmere which was what 95C (I honestly don't remember)?

 

As to the rest, I'm sure the fans run louder, the heat sink is also much more beefy though, and even when you consider the odd-ass temp targets given for AMD cpu's it looks like their stock cooling gives a bit more freedom than your average i7 stock cooler (again not counting the 4770k which was atrocious.) 

[TheProfosist]

I think the Opteron only goes to 8 modules, or 16 "cores".

So that hasn't improved either. I know some IT guys that were pumped when the 16 core's launched.

AMD also uses fake temp sensors, supposedly reporting god only knows what temperature, where I have no fucking clue what their actual Tj max is. Since both are silicon and honestly material properties don't vary that much at the same lithography, I would expect Tj max of similar to Westmere which was what 95C (I honestly don't remember)?

As to the rest, I'm sure the fans run louder, the heat sink is also much more beefy though, and even when you consider the odd-ass temp targets given for AMD cpu's it looks like their stock cooling gives a bit more freedom than your average i7 stock cooler (again not counting the 4770k which was atrocious.)

From my understanding it's not as good materials and not as good of a process that leads them to not be able to take as much heat. Who knows though cause Intel does kinda fib on their nm measurements.