• 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 Bulldozer Finally!

drunkenmaster said:
It won't have half the FP throughput, if you are only using 4 cores or two modules, thats because you aren't using more than 4 threads anyway, in which case, you can't use the other fp, if you have 8 threads, then all 8 cores will be used anyway.
Click to expand...

Bulldozer is a half/shared FPU architecture when comparing it to the number of arithmetic cores that it has. So, if only two modules are active when boosted to 4.1GHz you get 2x FPUs and 4x AUs at that max turbo.

drunkenmaster said:
As for whoever suggested overclocking only one core per module, not going to happen, power gating works that each module can be shut down completely, not each core, so 2 modules using all 4 cores will use a LOT less power than using 1 core in all 4 modules.
Click to expand...

That depends on how granular the power gating is, but I get your point because of all the shared resources for the dual AUs.

As for turbo, most people will be overclocking anyway, the interesting thing will be if turbo works well for overclocking.

drunkenmaster said:
In some situations with insanely heavy workloads that use every last bit of the core you won't get 3.6Ghz(for instance) on all 8 cores.
Click to expand...

It depends on the TDP calculations they use, I should think that it's realistic workload related rather than maximal TDP, but even so, that should cover the CPU to use all 8 AUs at 3.6GHz.
 
Bleek said:
Do you have a link?
Click to expand...

I can't link to it as its on a competitors site.
It's VK, the ex Asus rep.
I'll quote
I've replaced two words with XXXX.

-VK- said:
To be honest...all these benchmarks people are posting are a load of rubbish...until you see something from "XXXX" in "XXXX" forums or an "official" AMD announcement, I'd probably just ignore them.

It's coming and whilst I can't say anything about it, I can say that I'm excited.
Click to expand...
 
Other than benchmarking I don't think i've ever seen all 4 of my cores at 100% usage, Im sure for most users this is also true. It's certainly a good idea, while it might not keep the ultimate overclockers happy, it will certainly be good news for most gamers. Power saving is starting to become a relevant issue as all these power hungry pc components keep getting added to our systems, to know that your processor is only drawing power when it really needs to can surely only be a good thing. Its the next logical step from cool n quiet or speedstep.
Do all these switching speeds etc apply to the BD in all conditions or simply when running at stock, will there be an option to turn it off like there is with cool n quiet?
 
personally when im video editing or converting formats ive seen steady 100% usage on my 1055T, but for gaming not a chance, although id love a spin on one that would cause that amount of cpu sweat
 
I'm interested in Bulldozer but running a Phenom II Tri-core at 3.5GHz it does make me question whether I actually need to upgrade or not?
 
I believe the NDA is 10 days prior to release so you should expect to hear anything until the second week of September.
 
10 Days prior to release is quite....abnormal is it not?

Surely there are chips which represent the retail version now and it would be in AMDs best interest to start trying to secure some sales with some data?

Is it a case of "It's not going to blow current stuff out of the water so let's hold the NDA up till right before release, at which point if we match current other chips people have waited long enough so they will wait 10 days more anyway" ?

None of us know for sure the reasons why NDA is still in effect but IF the Sept 19th date is accurate, even if it's a soft launch why the NDA still? Could be nothing at all, could be indicative of another delay, could be indicative of Board/BIOS bugs that need ironing out and the chip only performs well in AMD test nominated systems?

Who knows, but seems strange!
 
Last edited:
I thought AMD's NDA's usually end on the release date? As it saves rewriting them every time there is a delay etc, its nothing abnormal, it just seems the way AMD works.
But I do see your point in releasing decent figures to secure sales etc, it is very quiet if the supposed date is correct.
 
yes seems strange to me which is why i asked..

wasnt sure on these things but if september 19th is correct i would have thought someone would be testing releasing figures by now. nothing at all in the magazines.
which considering are usually a month behind means no figures from them until after release which seems very odd..
 
a. None of the benchmarks that have been floated out so far are representative of actual performance. Unless you see something OFFICIAL from AMD, don't believe it.

b. To respond to the FP throughput, I am guessing that this group does very little FP work and that the "half the FP throughput" is part of the FUD that I hear being pushed around by other chip companies.

I recommend that people post up what apps they are talking about and identify whether or not they are going to be compiled to support 256-bit FP or whether they will continue to run SSE 128, or at best maybe AVX-128.

The only people that are going to use AVX-256 are probably people in the HPC market.

As a matter of fact, heavy FP would prevent you from doing a lot of overclocking because it eats a ton of your power budget. So if you are seeing lots of overclocking headroom, then don't worry about FP.
 
You must log in or register to reply here.
Back
Top Bottom