Here's a little teaser....

[PhilDoc]

no, you didn't, you just happenned to post then and got caught in the message..No issue with you here..

Glad to hear it and thanks. So, back to business and say a prayer for your friends for me.

[akaBruno]

Yup, everyone needs to take a deep breath and calm down. Not everyone's a mudderfudder just because they question something about AMD's marketing strategy.

I just hope that s7e9h3n gets his mobo's mojo workin.

And... a prayer for your friend's and their families MOVIEMAN.

BRUNO

[tsuehpsyde]

What's up with the clockspeeds Stephen?

And Dave, thanks for the info. Don't mind the people startin' stuff.

[akaBruno]

Those multipliers must have to be set separately, for each chip.

[PhilDoc]

I'm suspecting a bios or mb issue, but that would just be a quess.

[gallag]

god, you're a tool

If you disagree with these people thats your right but to attack them is not only wrong it's bad manners.

First movieman,Did you see the part where I posted my opinion and dave called me a tool, I think you are one of the best mods about here, but this is disappointing

Dave, the main thing that annoyed me about chatting over the Internet is that people act different that if you were chatting in a bar etc because they are not physically near you. What I mean is, If you were facing me you probably would not be so quick to call me names. I like moviemans saying "attack the argument not the person" even if it dosent apply to all.

[s7e9h3n]

Sorry I haven't been able to update this thread today, I've been a bit busy. Here's the deal in regards to the B1 vs. BA fiasco:

B1 is the silicon upon which ALL current Barcelona's are built and BA is a REVISION of the B1 silicon. So in essance, both sides were correct in their claims. To put it in more familiar terms for those who've had experience with the last generation of AMD's - consider BA a revision of the core/imc just like "BC", "BH", "BN", etc. with the FX57's. When AMD starts ramping up core speeds, we will then see the next generation of silicon -> B2.

Oh, and it seems as if I got a broken cpu. There's a identical system running in the lab and both procs properly boot @ 2g. I'm now waiting for a replacement chip - and possibly a few others - to arrive, hopefully, in the next few days....this sux. I'll see if there's any way to get this thing to run properly, but don't hold your breath :p

[Kenetixx]

Leave us hanging s7e9h3n why dont you :P
Bad luck on the broken chip, i seen abotu 20 of today at work , wish i could have nabbed a few .

[kyosen]

Many thanks to s7e9h3n, Nedjo, dave_graham, for clarification about stepping/revision!!!

[zakelwe]

there's a difference between qualifying a platform and shipping for revenue. that gap can be SIGNIFICANT in time. I'm not shocked at all. I know quite a few folks that are qualifying their products on pre-release intel silicon.
cheers,

dave

According to AMD in late June they would be shipping for revenue in August.

http://www.amd.com/us-en/Corporate/V...118193,00.html

They then confirmed this on the very last day of August

http://www.xbitlabs.com/news/cpu/dis...831115038.html

So, assuming that those are all BA revision of B1 stepping as S7e9h3n says then I think it likely that Anand was also sent a BA rev step B1 so the values they got reflect the current production K10.

Your claim that Anand is not a big fish, so to speak, might be true. However they were big enough for AMD to send them a B2 as well. Would they send them a B2 and not a BA?... hmm, highly unlikely. More likely is that your field engineer does not know and has fed you duff information which you are trying to reconcile assuming he/she to be correct.

Regards

Andy

[alayashu]

another teaser
http://youtube.com/watch?v=F7LNUkHa7U8

that's 8x8 Skulltrail hehe)





muhahahha

[Kenetixx]

I just wasted abotu 40 second sof my life i will never get back on that crap

[KTE]

Any latest scores here?

If you guys can show the highs and lows of the processor when you get time (thanks), I'm just interested in the raw architectural power and weaknesses based on the specifications I know of it. AMD engineers seem to think its a major overhaul and are happy with it, and they aint no dummies but pioneers.

One thing: I 'aint a corporation or enterprise worker dealing with database centers and tons of servers (Uni student yet but do work with them at hospital), but I clearly understand the difference in server benchmarking and desktop: two different ball games. They are unequivocally very different to one another, as servers usually tend to perform very differently based on application and software optimization (in-house usually). I know it from the medical center I work in part-time for experience which is a sub-department of the government, and they have their own engineers, analysis, system setups and software which they base their analysis and reports on.

Just hope you can sort out this core revision mess once and for all -> reconcile all the info spreading around rather than leaving it for 3rd 4th 5th parties to make sense of, which they can never (missing common sense you know ) and then it all turns haywire into a pig farm.

Movieman, bro, thanks for keeping it a little sane and yeah, I realize reality from fanfare very well. People die all the time every second, some more fortunate with better lives than most others in non-Occidental nations. ANY loss is always unfortunate though. My heartfelt condolences to the bereaved.

K10: Last I heard (well I sent buck loads of emails ya know ) from my uncle whose worked in this enterprise business for long so is in direct contact by technology firm reps on all sides where need be especially second quarter where his company departments usually upgrade systems (though I get no answers ), the rep dealing with him at AMD was saying its upcoming chips outperform Intels for general business, high performance, Int and FP intensive computing (yes all) based on price/performance, performance/watt, and they will have 2.5GHz delivery parts by end of year (Opteron - which is only what his corporation and their affiliates wanted). This is roughly what Dell reps said to him aswell.

Dave, Steve or anyone "informed" know if there's any truth in that?

[akaBruno]

Oh, and it seems as if I got a broken cpu. There's a identical system running in the lab and both procs properly boot @ 2g. I'm now waiting for a replacement chip - and possibly a few others - to arrive, hopefully, in the next few days....this sux. I'll see if there's any way to get this thing to run properly, but don't hold your breath :p

Is there anyway to force the multiplier to a Manual setting in the bios?

[dave_graham]

First movieman,Did you see the part where I posted my opinion and dave called me a tool, I think you are one of the best mods about here, but this is disappointing

Dave, the main thing that annoyed me about chatting over the Internet is that people act different that if you were chatting in a bar etc because they are not physically near you. What I mean is, If you were facing me you probably would not be so quick to call me names. I like moviemans saying "attack the argument not the person" even if it dosent apply to all.

i apologize. was again getting frustrated at being attacked in character when, as you so succintly say above, you'd probably never say the same thing in close proximity.

that being said, beers on me in Boston if/when you come up.

dave

[dave_graham]

According to AMD in late June they would be shipping for revenue in August.

http://www.amd.com/us-en/Corporate/V...118193,00.html

They then confirmed this on the very last day of August

http://www.xbitlabs.com/news/cpu/dis...831115038.html

well, they're shipping for revenue (according to them) but not to the channel. this is for direct buy enterprise. they didn't make the august rev. date but they will make Q3.

So, assuming that those are all BA revision of B1 stepping as S7e9h3n says then I think it likely that Anand was also sent a BA rev step B1 so the values they got reflect the current production K10.

i still maintiain that's not the case. but, we'll see. between S7 and I we should be able to validate those #s...which is the sign of good science anyway.

Your claim that Anand is not a big fish, so to speak, might be true. However they were big enough for AMD to send them a B2 as well. Would they send them a B2 and not a BA?... hmm, highly unlikely. More likely is that your field engineer does not know and has fed you duff information which you are trying to reconcile assuming he/she to be correct.

Regards

Andy

anand IS a big fish in the prosumer market but that's about where it ends. they DID send him B2 procs....those 2.5ghz ones with the Sharpie writing on the top . that's the coup d'etat in the whole deal...they're showing that:
a.) they CAN scale quickly to higher speeds
b.) that Barcelona really has a lot more breathing room than was previously thought.

again, it's the double-edged sword i talked about in earlier posts....

as for my FAE, well, he's been @ AMD for a LONG time and has no compunctions whatsoever in telling me what parts of AMD function well and those which don't. that's the screening question for finding out if someone's telling the truth or not. And, he comes highly recommended from several of my other AMD collateral in the area (folks at the Boston Design Center) as being a no-holds-barred kinda guy....

cheers,

dave

[dave_graham]

Is there anyway to force the multiplier to a Manual setting in the bios?

no at least on the boards i've dealt with.


dave

[metro.cl]

well, they're shipping for revenue (according to them) but not to the channel. this is for direct buy enterprise. they didn't make the august rev. date but they will make Q3.



i still maintiain that's not the case. but, we'll see. between S7 and I we should be able to validate those #s...which is the sign of good science anyway.



anand IS a big fish in the prosumer market but that's about where it ends. they DID send him B2 procs....those 2.5ghz ones with the Sharpie writing on the top . that's the coup d'etat in the whole deal...they're showing that:
a.) they CAN scale quickly to higher speeds
b.) that Barcelona really has a lot more breathing room than was previously thought.

again, it's the double-edged sword i talked about in earlier posts....

as for my FAE, well, he's been @ AMD for a LONG time and has no compunctions whatsoever in telling me what parts of AMD function well and those which don't. that's the screening question for finding out if someone's telling the truth or not. And, he comes highly recommended from several of my other AMD collateral in the area (folks at the Boston Design Center) as being a no-holds-barred kinda guy....

cheers,

dave

Dave somhow you are loosing credibiity in my book.

Check this:

“AMD today shows the path to better performing design for demanding data center requirements, again. In fact, here is a proof to you of our 2.50GHz [quad-core] processor that is going to be available in December and what it is going to deliver to the market. […] This chart shows incremental performance that is going to be delivered on top of that existing [2.0GHz] product: you can see 15% performance benefit, 15% performance increase that will be delivered [by] our 2.50GHz product,” said Randy Allen.

So phenom at 2.5Ghz will only be 15% faster than Phenom at 2.0GHz so much for the incremental speed gains or linear speed gains.

Source

[Motiv]

So phenom at 2.5Ghz will only be 15% faster than Phenom at 2.0GHz so much for the incremental speed gains or linear speed gains.

Source

I don't understand the Quote at the end.

AMD Phenom FX-80 quad-core processor is projected to operate at 2.20GHz – 2.40GHz clock-speed, have 2MB of L2 cache (1MB per core), 2MB shared L3 cache, dual-channel PC2-8500 (DDR2 1066MHz) memory controller and AM2+ form-factor with HyperTransport 3.0 bus support. The AMD Phenom FX-90 and FX-91 models are expected to operate in 2.20GHz – 2.60GHz clock-speed ranges and support dual-processor configurations.

The figures they show don't add up. Either it's a dual core, with 1MB cache or it's a quad with 512.

[BeardyMan]

metrocl, what were you expecting?35%

[metro.cl]

metrocl, what were you expecting?35%

dude learn to read i always said no linear not incremental.

Others kept saying wait and see more MHz will do wonders, is the stepping, etc..., that is a lie!

[PhilDoc]

Well, going from a 2.0G to a 2.5G you would expect a 25% increase if it scaled perfectly. If you look at the Anand Phenom article it fell short of that but not by much. When they compaired the two only on two of the benchmark were at the 15% range, two others were at the 25% range. The average increase was 19%. Of course we don't know the stepping of the 2G chip, but as Dave said the 2.5G look to be a B2 stepping according to cpu-z. Looks pretty good to me.

It looks like Anand has done a little revision of their Quad Barcelona article. They've added the prices and looks like they've revised some of the benchmarks. In the end it look as though Barcelona does very well clock for clock against the Intels. Only losing in two tests that are biased toward the Intels to begin with. In the other benches the Barcelona is quicker clock for clock. Considering how immature all of this is, imho this bodes very well for Barcelona.

As for the Xbitlabs article, don't you just love the Intel Add next to a AMD article. I just about rolled on the floor when I saw that. I think when Allen said 15% he was being concervative

[Motiv]

Well, going from a 2.0G to a 2.5G you would expect a 25% increase if it scaled perfectly. If you look at the Anand Phenom article it fell short of that but not by much. When they compaired the two only on two of the benchmark were at the 15% range, two others were at the 25% range. The average increase was 19%. Of course we don't know the stepping of the 2G chip, but as Dave said the 2.5G look to be a B2 stepping according to cpu-z. Looks pretty good to me.

The 2.0G was B1 stepping.

Now if there is no difference in the steppings, then yes. Perhaps 15% is a conservative estimate but 20% on average would be max.

It looks like 15% is going to hold true.

[PhilDoc]

The 2.0G was B1 stepping.

Now if there is no difference in the steppings, then yes. Perhaps 15% is a conservative estimate but 20% on average would be max.

It looks like 15% is going to hold true.

Well its early yet, but it looks to me from reading that article. 15% would be the least you'd get. But you're right, if that was a B1 stepping, and I believe Dave that it was, then we just don't know and 15% maybe right. Either way that's not bad. Increasing clock speed rarely gives you a 1/1 increase

[zakelwe]

i still maintiain that's not the case. but, we'll see. between S7 and I we should be able to validate those #s...which is the sign of good science anyway.

Yes indeed, what we need is a greater mass of evidence to be able to judge fully, so both your results will be appreciated. From a home user perspective it is a shame we did not see Phenom FX being released today being benched on multiple sites with it's relevant memory and motherboards, to give a better representation to the enthusiast. Opteron K10 looks pretty good for server applications in it's price range, but that does not really interest me.

Regards

Andy