I wouldn't expect a giveaway to include a CPU released publicly before the NDA is up. Having a random guy sign an NDA would be a bit strange as part of a prize, but maybe they will. I just wouldn't expect it.
Printable View
I wouldn't expect a giveaway to include a CPU released publicly before the NDA is up. Having a random guy sign an NDA would be a bit strange as part of a prize, but maybe they will. I just wouldn't expect it.
Quote:
Originally Posted by informal
That's not FX PCs, I said that because FlanK3r pics + comment was misleading ;)Quote:
Originally Posted by FlanK3r
hmmmm looks promising
better ready the 75" screen :D
how much cpu power is needed to play dirt 3 in eyefinity? i dont imaging much, so i think they could have done better demos
Ok, 2 PCs, but second photo is Think the right photo :)
Well, it's multithreaded, got decent FPS at extreme resolutions (something you need a good CPU for to reduce bottleneck of multi-GPU) and probably used a large amount of ram.Quote:
Originally Posted by Manicdan
No, it's not a synthetic test, but perfect real-world gaming situation. (If you can call eyefinity real world)
AMD readies Bulldozers to ship next month
woot! if true
http://www.theregister.co.uk/2011/07..._2011_numbers/
In your haste to post, you should of added this :pQuote:
Originally Posted by duron
El Reg discovered that AMD's top brass meant to say that these two chips would ship for revenue next month, but it is still not clear when they will be formally announced.
Dont know if thats any kind of news.But from the article:
Thats pretty much exactly increase in core count.12+35%=16.2.Quote:
On the server front, Siefert said that the 16-core "Interlagos" Opteron 6200 processor, the first server chip to make use of the "Bulldozer" core and implemented in GlobalFoundries' 32 nanometer processes, would ship this quarter to OEM partners. He added that the chip would offer about a 35 per cent performance boost compared to the current 12-core Opteron 6100s.
If that is the case, then that would be horrific for AMD. They would need these things to clock like no tomorrow.Quote:
Originally Posted by RaV[666]
That is some pretty bad news if true... either clock well or low TDP.
Or he might be giving a realistic view of performance increase expecting software that his clients run to not scale linearly with core count. Amdahl's law and all that.
Unlikely, but hey.
at best one could interpret performance increase per core :p
In the CC Seifert talked about 35% more performance in the context of HPC.
Was that solely for software using AVX or in older HPC suits.Quote:
Originally Posted by Dresdenboy
If heavy FP code can be up to 35% faster on Interlagos it's quite good actually.
HPC suffered with memory bandwidth more than pure computing ability.Quote:
Originally Posted by Lightman
Correct! Still this is same G34 platform with 4 64bit channels and Interlagos officially supports up to DDR3 1600 which is 20% improvement from DDR3 1333.Quote:
Originally Posted by undone
Finally the performance of Bulldozer ES FX-8150 from OBRovsky:
Attachment 118100
Own AMD shares? Sell Sell Sell.... :shakes:
(but dont bite at that comment! I know it is ES and I also have a huge amount of hope and faith in bulldozer, I cant see that it can be as bad as in the graph above)
Edit:
At a second glance, that graph suggests that if that 8150 was clocked the same as the 1100T, the results would be even closer between Thuban and Zambezi clock for clock.
When will we have some actual figures of non crippled chips!?
Dont panic james.Thats OBR, he hates AMD.There are benchmarks in there purposefully put there to lower AMD scores, im sure of it.
We know for example that AES hardware acceleration in BD is much worse than in sandy, which will drown mean score.
Of course he could test more demanding encryption scheme like aes-twofish , but he most probably didnt :).If AMD is targeting its price at 300$, it cant be worse than 2500K.
As long as he don't tell us what mb+bios (agesa microcode) and chipset drivers, memory speed + OS etc etc he testing on. + it's a ES cpu. And how did he calculate the end score? (i see SuperPi in there among others *ouch*) :down:
You can safely just ignore it ;)
Remember this post from JF-AMD
OBR doesn't hate AMD, he just likes taking advantage of others and messing with people. It probably wouldn't make a difference if it were any particular company; an ass is an ass is an ass.
i see cpuz screen shot 4.2GHz @ 1.39xxx V
stablitiy unknown