Plausible speculation re: bulldozer ES performance:
http://crazyworldofchips.blogspot.com/2011/06/zambezi-es-performance-weirdness.html
http://crazyworldofchips.blogspot.com/2011/06/zambezi-es-performance-weirdness.html
Oh dear. This doesn't sound particularly positive. The talk of engineers leaving frustrated reminds me of Matrox getting picked clean by ATI and NV.
AMD Insiders Speak Out: BAPCo Exit is An Excuse for Poor Bulldozer Performance
http://www.brightsideofnews.com/new...-for-poor-bulldozer-performance.aspx?pageid=1
AMD Insiders Speak Out: BAPCo Exit is An Excuse for Poor Bulldozer Performance
http://www.brightsideofnews.com/new...-for-poor-bulldozer-performance.aspx?pageid=1
DarthShader
Regular
Yeah and nVidia left because of poor GPUs performance compared to Intels IGPs in SYSmark. 
AMD, nVidia and ARM should produce a compelling alternative if Sysmark isn't a representative benchmark in their eyes.
With respect to reasons for leaving, even if Sysmark totally ignores graphics performance AMD wouldn't be throwing a fit if Bulldozer could stand on its CPU performance. nVidia actually has a much better reason for leaving since they no longer make x86 chipsets and have no stake in that game.
With respect to reasons for leaving, even if Sysmark totally ignores graphics performance AMD wouldn't be throwing a fit if Bulldozer could stand on its CPU performance. nVidia actually has a much better reason for leaving since they no longer make x86 chipsets and have no stake in that game.
ARM doesn't allow independent and credible benches to come out in the open.
Besides, why is VIA leaving?
Probably they too have little or no benefit from competing in this benchmark, so they just used the case to gang up with the majority on the way out.
Bright Side of News, eh?
Well, we know that's as reputable as Fudzilla/Semi-Accurate/The Inquirer.
Last edited by a moderator:
I don't know, but why distract from the more important question. Why is AMD leaving? If Bulldozer was the bees knees they would have no reason to throw up such a stink. They could've protested the lack of GPU testing privately without creating this public spectacle.
I think we'll have the answer once Bulldozer launches. Sysmark isn't the only benchmark out there.
Not sure we can infer much about BD from the Sysmark thing (and Theo being adamant about something makes me doubly doubt that thing!). Considering they're in a tantrum over GPU stuff, I'd think it has more to do with Llano - it's likely that OEMs that are arguably more important for Llano's success run lotsa Sysmark (maybe? is Sysmark relevant to anything? I have no idea!), and Llano is an absolute dog in terms of CPU throughput. So it's probable they're pissed off that it looks awful there, and can't save it on the wings of its GPU. Then again, I may be wrong.
hoom
Veteran
I dunno what to think about the Sysmark thing but what a great time to leak an overclock score on presumably one of few samples that runs well
http://www.fudzilla.com/processors/item/23210-eight-core-bulldozer-overclocked-to-463ghz
I have no idea if thats a good super-pi score?
http://www.fudzilla.com/processors/item/23210-eight-core-bulldozer-overclocked-to-463ghz
I have no idea if thats a good super-pi score?
http://obrovsky.blogspot.com/
Original image has been cencored but it's visible that it has taken at minimum 10s to compute it. IIRC I7 breaks 10s at around 4GHz or so.
Original image has been cencored but it's visible that it has taken at minimum 10s to compute it. IIRC I7 breaks 10s at around 4GHz or so.
SNB should do ~8 sec. on this clock-rate.
My dual core Wolfdale 3.2GHz is ready in 14.697s. Not bad for just 2 cores without some fancy hyperthreading.
Probably SuperPi likes caches, as wolfdale has 3MB L2 for each core.
My dual core Wolfdale 3.2GHz is ready in 14.697s. Not bad for just 2 cores without some fancy hyperthreading.
Which is irrelevant considering Super Pi is not multithreaded