TLB bug still present in B3 rev Phenoms

[Nedjo]

Theinq also confirmed it to be 50% faster than Core 2 and doing 30K 3Dmark06. Not even to mention the "dancing in the aisle".

K10 is more then 50% faster then Conroe in some bandwidth heavy tasks, and in CF-X setup K10 can achieve 30K in 3DM06!

[SEA]

They don't mention the TLB bug specifically but this is the one they're talking about anyway.

Don't know about 9550 and 9650 being pushed back.

Now there is another French site saying they contacted AMD France to clarify the matter and have been officially told that there actually were no problem and that 9700 and 9900 had started to get in big OEMs hands, that they were ramping up the delivery and that everything was still on track for a worlwide launch at the end of this quarter.

http://www.presence-pc.com/actualite...-retard-27203/

Wait a minute. Hmm right. I checked that article (that one about delay) again - they say nothing about 9700 and 9900. They only say about 9550 and 9650 being pushed back... So nobody ever said about delay of 9700 and 9900 which are most awaited.

[AbelJemka]

http://www.elitebastards.com/cms/
B3 = no bug

[Shintai]

K10 is more then 50% faster then Conroe in some bandwidth heavy tasks, and in CF-X setup K10 can achieve 30K in 3DM06!

[LowRun]

Wait a minute. Hmm right. I checked that article (that one about delay) again - they say nothing about 9700 and 9900. They only say about 9550 and 9650 being pushed back... So nobody ever said about delay of 9700 and 9900 which are most awaited.

You're right about the initial article but unfortunately yes somebody did say something about a 9700/9900 delay.

http://arstechnica.com/news.ars/post...r-variant.html

.

[SEA]

This sounds even more FUD...
And other sites refer it as "as undisclosed AMD representatives said in an interview with Ars Technica web-site that the new Phenom chips are now only scheduled for the introduction in Q2 2008".

And i cannot understand: they (amd) have B3 and it is ok and this was confirmed by AMD and we even know name: Pat Moorehead; So why would they delay 9700? arstechnica says because they work on 9100E. I don't beleive that they only can work on one model one time...

[Cooper]

This sounds even more FUD...
And other sites refer it as "as undisclosed AMD representatives said in an interview with Ars Technica web-site that the new Phenom chips are now only scheduled for the introduction in Q2 2008".

And i cannot understand: they (amd) have B3 and it is ok and this was confirmed by AMD and we even know name: Pat Moorehead; So why would they delay 9700? arstechnica says because they work on 9100E. I don't beleive that they only can work on one model one time...

Do you know how long it takes to make at least couple dozen thousands CPUs ?

[SEA]

Sure. The whole production cycle is alsmost 3 months.
That's exactly what i was trying to say: In december AMD said that they will have B3 samples out at the end of january. And thus those "early B3 samples" mentioned by arstechnica are pre-B3.

[justapost]

Who cares above 2100Mhz right now? It doesnt even work at 1800 or 2000Mhz. The only thing that matters is the core speed for the bug.

I talked about phenom stability. The Nb speed has an valuable impact on the memory bandwith and latency. The Core can be overclocked more with an nb speed below 2100MHz.

Also they can not test it as such.

I asked for errata occurencies without virtualisation on the x86-64 mailing list.
There is one person who had an lockup after 1-2 days crunching. He did not have a lockup with the tlb-workaround applied.
He could reproduce the lockup so that seems to be a good testcase.

Made a kernel whom disables the bios fix without applying the workaround. Will try to reproduce his results.
1-2 days crunching ain't no quick test case. I'm curious if i can reproduce the lockup.
If so i can try the same stuff with watercooling and/or overclocking. If it's a hotspot problem the lockup should occure earlier on an overlocked system and later with watercooling.

[JumpingJack]

K10 is more then 50% faster then Conroe in some bandwidth heavy tasks, and in CF-X setup K10 can achieve 30K in 3DM06!

I don't want to be mean ... but you could provide links showing the data... I can handle the 50% in various configurations, it should really perform there... but 30K in 3DM06.... as I recall, even with a 5GHz+ Yorky and xFire that pinnacle has not been breached.

[STaRGaZeR]

I don't want to be mean ... but you could provide links showing the data... I can handle the 50% in various configurations, it should really perform there... but 30K in 3DM06.... as I recall, even with a 5GHz+ Yorky and xFire that pinnacle has not been breached.

Be careful, he has said CF-X, now he will say the 30k 3DM06 is with four 3870s, and then think he has pwnt you