AMD Zambezi news, info, fans !

[muziqaz]

JF-AMD http://www.overclock.net/amd-cpus/11...aunch-faq.html. Perhaps it just me, but this guy comes across as arrogant and sarcastic.

It is just you

[Brother Esau]

So the mythical Unicorn has still not appeared huh?

Man, I tell you this is beyond ridiculous! 4 years now and still no Bulldozer WTF AMD! Im a big AMD fan as most of you know but I could not wait any longer 3 months ago when my server needed upgrading due to parts failure. Come back 3 1/2 months later and still no sign of these darn chips anywhere......why?

B.T.W ...who's bright idea was it to redo this forum in this ghastly face lift and IN non functioning manner....yet another BIG....W.T.F!!!

[muziqaz]

<..>

B.T.W ...who's bright idea was it to redo this forum in this ghastly face lift and IN non functioning manner....yet another BIG....W.T.F!!!

At least you have your nickname in red. It is almost impossible to read other peoples nicknames in this forum color. dark grey nicknames, grey forum background :/

[-Boris-]

These quotes from JF's FAQ seems to confirm my anticipations, or lack of there of:

Q. Is IPC higher on bulldozer? All I care about is IPC.A. IPC is simply a measurement. What if IPC was 2X what it is today, but clock speed was 500MHz. Is that what you want? You are getting double IPC, right? IPC is only one measure. The people that are telling you IPC is the only thing that matters have an agenda. Taking only one measurement, out of context, is like trying to say that a person's weight is all that matters. I weigh 195. Does that make me fat? Does that make me skinny? It is impossible to say unless you know my height. IPC is like weight - it tells you something in context to other factors, but is meaningless on its own.

Q. What about single threaded performance?A. See above. Also, if all you care about is single threaded performance, might I recommend a lovely, inexpensive single core processor for your system?

The way he try to mock people who care about IPC and single thread performance makes me believe he is trying to justify bad performance in these areas. We know frequency range, so IPC is important! And single thread performance is important to everyone who isn't building a server. It doesn't matter if you have a million cores, if single thread performance matches a K6-2 you won't be able to play any games today, just using windows would suck

IPC will be low, maybe lower than Phenom II, and that goes for single thread performance as well. Their recent statement of 35% more performance with 33% more cores was the first clear warning of this, that's down from official numbers of 50% more throughput last year. I haven't seen any trustworthy benchmarks yet so this is all speculation, but since all the fuzz is about Llano it doesn't look like BD is the winner in AMDs portfolio. I mean even Llano was a disappointment, you can get more gaming performance at the same price with an AM3 board, Athlon II and a cheap graphics card. Llano is good for laptops and ITX systems, but if that's all AMD will be good at then it looks bleak.


By the way, I'm old AMD fanboy, and i guess I will buy BD after all since i think competition is important, but that's not the best of reasons to switch to the latest CPU.

[demonkevy666]

These quotes from JF's FAQ seems to confirm my anticipations or lack of there of:



IPC will be low, maybe lower than Phenom II, and that goes for single thread performance as well. Their recent statement of 35% more performance with 33% more cores was the first clear warning of this.

I'm calling BS on that.
there was no mention of core count from amd, when that was said.

[-Boris-]

I'm calling BS on that.
there was no mention of core count from amd, when that was said.

No, they just said that the upcoming Opterons would be 35% faster than the current ones. And that's on similar, or a tad higher frequencies. And as we all know upcoming Opterons has 33% more cores. And of course they are comparing top models and not some lower end BD. If I say my new car is 20% faster than the old one I mean that my new car does 240Kph instead of 200Kph, not that it actually does 300Kph but I was only driving 240Kph when I made the comparison.

[demonkevy666]

No, they just said that the upcoming Opterons would be 35% faster than the current ones. And that's on similar, or a tad higher frequencies. And as we all know upcoming Opterons has 33% more cores. And of course they are comparing top models and not some lower end BD. If I say my new car is 20% faster than the old one I mean that my new car does 240Kph instead of 200Kph, not that it actually does 300Kph but I was only driving 240Kph when I made the comparison.

was not a comparison to single chip but whole family and whole family ranges form 4-16 core.

[Olivon]

No problem found so far with the Bulldozer OC Contest currently in progress.

http://www.xtremesystems.org/forums/...=1#post4941471

[-Boris-]

was not a comparison to single chip but whole family and whole family ranges form 4-16 core.

I saw this answer coming, that's why I made the car analogy. In this business, all companies, has always talked about their fastest product. There is simply no sense in making claims based on average performance for an entire line of products. What if they decide to release a lot of models with 4 cores, is it only 5% faster? Or if they only release on single version with 8 cores is it 50% faster? If they compared the way you're saying then the result would differ greatly depending on how many low end models they intend to release. If intel releases a lot of different ULV models i7 around 1.5GHz, would that suddenly make Phenom II faster than i7? No of course not, Phenom II would still get it's ass kicked then you compare the high end models.

It's never been compared that way before, what's the reason they would do that now?

[2good4you]

IPC will be higher, which also JF stated earlier, but not by how much, and that you'll see on launch!

[LightSpeed]

I think even if IPC is the same while clocking a ghz higher and having four more cores, its not that bad. I still think it will be higher though.

I agree, the recent comments havent really been that encouraging, but im still eagerly awaiting the first real leaks and products to hit shelves.

[PerryR]

The way he try to mock people who care about IPC and single thread performance makes me believe he is trying to justify bad performance in these areas.

I thought he was providing a response to those that think if single-threaded performance isn't high enough in comparison to Sandy, then it's a failed product; irregardless of how it performs in other areas. I've seen people, on various forums, make such a claim.

[BeepBeep2]

was not a comparison to single chip but whole family and whole family ranges form 4-16 core.

Even if that was 12 to 16 core comparison, IPC will be higher 8 cores and less.

CMT is CMT for a reason, AMD compared it to HyperThreading for a reason, and I believe chew* was right when he said desktop parts will function more like a 4 core that can execute 8 threads.

In multi-thread, I actually expect Thuban level performance from well coded programs.

[zir_blazer]

IPC will be higher, which also JF stated earlier, but not by how much, and that you'll see on launch!

At first glance the longer pipeline and higher Frequencies makes Bulldozer sound like a NetBurst inspired thing with lower IPC that current K10 that could be as big as a fail as Prescott was, hot and ugly. If wasn't because JF-AMD insistense on the "IPC increases" topic (Refer to Hans de Vries signature ), I would be dubious about what the hell Bulldozer is. But that very JF-AMD statement, plus the fact that the current trend on performance per Watt means that no one sane enough would design anything like Prescott, makes me believe that Bulldozer could end being quite... Interesing.
At the bare least you have a bit higher performance per clock, and you also have a bit higher Frequencies. Even if it doesn't beat Sandy Bridge it will come dangerously close.

[freeloader]

http://translate.googleusercontent.c...Twc6Msj2d0jwcA
Don't know if it's been posted yet.

[PerryR]

http://translate.googleusercontent.c...Twc6Msj2d0jwcA
Don't know if it's been posted yet.

It was found that the scores were copy and pasted from another article on that site:

http://www.sisoftware.net/?d=qa&f=cpu_intel_sb_turbo

http://hardforum.com/showpost.php?p=...8&postcount=15

As such, it's likely 99% fake.

[JumpingJack]

It was found that the scores were copy and pasted from another article on that site:

http://www.sisoftware.net/?d=qa&f=cpu_intel_sb_turbo

http://hardforum.com/showpost.php?p=...8&postcount=15

As such, it's likely 99% fake.

I'm confused. While I agree, this looks dubious ... how does copy and pasting archived data from a prior article make it a likely fake? The scores for SB are taken from identically configured CPUs, i.e. it as C&Ped from a 3.0 GHz base CPU and quoted as such in the compare. Reviews often pool archived results otherwise it would be inefficient to re-run every bench for all comparable CPUs anytime a single new product is released.

[PerryR]

I'm confused. While I agree, this looks dubious ... how does copy and pasting archived data from a prior article make it a likely fake? The scores for SB are taken from identically configured CPUs, i.e. it as C&Ped from a 3.0 GHz base CPU and quoted as such in the compare. Reviews often pool archived results otherwise it would be inefficient to re-run every bench for all comparable CPUs anytime a single new product is released.

It seems the guy was incorrect about the article being faked, he changed his post:

http://hardforum.com/showpost.php?p=...8&postcount=15

LOL. Seems it's just copied data.

[JumpingJack]

It seems the guy was incorrect about the article being faked, he changed his post:

http://hardforum.com/showpost.php?p=...8&postcount=15

LOL. Seems it's just copied data.

Well, that does not mean that it is not fake either. A high profile product launch like this always generates wackos and dorks trying to reach across the interwebble to perform the cyber equivalent of a purple nurple. As a general rule, it is best to not lend much credence to anything 'leaked', the best data will come from reputable sites that have received actual shipping product to review in which case the data published will be known when all NDAs are lifted.

[Chrono Detector]

Would prefer to see an official date launch for AMD's Bulldozer because so far we haven't heard anything, and it sucks being kept in the dark for so long.

[massman]

Isn't it really simple? If AMD can come up with something that is competitive in some benchmarks, easier to bench and more fun to tweak ... I'll get one. If it's not competitive, a to bench and no tweaking options at all ... I'll probably still get one.

The IPC doesn't really matter for overclockers, I guess. In the end, it's all about the performance you can obtain with certain cooling. Why would you need to have superior IPC if you cannot get the clocks up. If you get 50k Vantage with Bulldozer at 6G or Sandy at 5G, but there's more room for frequency improvement on Bulldozer, you'd pick the Bulldozer.

When Deneb came out, it was competitive with Bloomfield in 05 and 06, but needed much higher clocks to do it. Who cares ... it was competitive

[Blacklash]

I can understand why AMD might not want to announce to the world an exact date like Sept 19th.

They want people buying as many of their other CPUs as possible. I've seen many folks buying "tide me over" CPUs in the past. AMD may want that to continue as long as possible. If people know an exact concrete launch day, I'd imagine AMD's other CPU sales may drop off rather well as said date approaches.

[liberato87]

I can understand why AMD might not want to announce to the world an exact date like Sept 19th.

They want people buying as many of their other CPUs as possible. I've seen many folks buying "tide me over" CPUs in the past. AMD may want that to continue as long as possible. If people know an exact concrete launch day, I'd imagine AMD's other CPU sales may drop off rather well as said date approaches.

that's what jf said and it is not hard to get.

[Hans de Vries]

http://translate.googleusercontent.c...Twc6Msj2d0jwcA
Don't know if it's been posted yet.

Well,

The latency results for simple sequental access show that prefetching is disabled in this ES....

In Sandy Bridge preftching reduces the L3 latency by a factor 2 or so. Latency reduction
is responsible for most of Sandy Bridge's IPC increase over Nehalem.

Regards, Hans

[FlanK3r]

Even if that was 12 to 16 core comparison, IPC will be higher 8 cores and less.

CMT is CMT for a reason, AMD compared it to HyperThreading for a reason, and I believe chew* was right when he said desktop parts will function more like a 4 core that can execute 8 threads.

In multi-thread, I actually expect Thuban level performance from well coded programs.

no sense...because 45nm Thuban with die size 380mm = 32nm 315mm FX? What about than at theoretical Thuban at 32nm with higher clocks and lower die size than FX? No, it must be better in good optimalized programs