Don't get fooled by .... nvidia attacking Ati
- Thread starter nagus
- Start date
Vortigern_red
Newcomer
I have, obviously, misunderstood all these comments about "the statement about the PCIe is true" because the statement as I read it is "ATI fails PCI express compliance" (word for word as NV wrote it) whereas most seem to be reading it as "ATI has not got PCIe compliance certification"
Obviously the latter is true WRT the X700 but I got the impression that people here knew that ATI had in fact taken and failed the tests with all the products not on the list. The Inq story presents no evidence (nor even claims) that ATI had failed the tests, it just poses the question.
My bad, anyway as I currently see it the statement is false based on the provided evidence (the Inq story). NV might actually know if it is correct of course, but be unable to provide the sources so posted the Inqs suggestion. But I am not going to rely on NVs word.
Actually, thinking about it the whole language on this thing is really sloppy. If they had said the X700 has no certification, fair enought. If they had presented real figures with Doom3 against the cards real competitor the 128MB X700XT, they would have still had a decent, and beliveable, and marketable lead, if they had said that the X700 is currently only available in a single GPU configuration it would have been true (assuming, as some here have suggested, that the RV410 is capable of multi GPU) I'm not even going to mention that even my old 486 is only 100% slower than the 6600GT in Doom3, its in the loft and for the timedemo I did not even plug it in!
.
AFAICS it makes any valid points in the whole thing near useless! A lot of people will dismiss the whole thing because of the outrageous claims. Talk about the boy who cried wolf!
Obviously the latter is true WRT the X700 but I got the impression that people here knew that ATI had in fact taken and failed the tests with all the products not on the list. The Inq story presents no evidence (nor even claims) that ATI had failed the tests, it just poses the question.
My bad, anyway as I currently see it the statement is false based on the provided evidence (the Inq story). NV might actually know if it is correct of course, but be unable to provide the sources so posted the Inqs suggestion. But I am not going to rely on NVs word.
Actually, thinking about it the whole language on this thing is really sloppy. If they had said the X700 has no certification, fair enought. If they had presented real figures with Doom3 against the cards real competitor the 128MB X700XT, they would have still had a decent, and beliveable, and marketable lead, if they had said that the X700 is currently only available in a single GPU configuration it would have been true (assuming, as some here have suggested, that the RV410 is capable of multi GPU) I'm not even going to mention that even my old 486 is only 100% slower than the 6600GT in Doom3, its in the loft and for the timedemo I did not even plug it in!
.AFAICS it makes any valid points in the whole thing near useless! A lot of people will dismiss the whole thing because of the outrageous claims. Talk about the boy who cried wolf!
WRT PCIe compliance, as you've probably seen ATI has X600 listed in the components section, whilst NVIDIA has lots of board variants listed in the add-in-boards section. ATI basically say the this is purely an issue of time and paperwork and they decided to validate things differently to NVIDIA.
For compliance you have a "plug fest" where, AFIAK, you basically try your components/boards out with a bunch of other pcie devices, then run through a bunch of other tests and then run up a load of paperwork based on that test. NVIDIA have decided to do this at the board level, so they have gone through that process with each of their boards to list each one with compliance - by putting X600 through as a "component" they are saying this covers the entire X600 line, as well as board vendor variants for that level of compliance (which, probably purposefully, inadvertently suggests that NVIDIA's vendors would need to go through their own validation for the same level of compliance if they are not using the reference platform from Flextronics [edit] It seems that a bunch of NVIDIA boards are now listed in the components section as well, which would negate that).
X600 is RV380, which is basically ATI's reference platform for PCIe so, unless there were implementation specific issues, its unlikely that any subsequent chip (which is all of them) is going to be less compliant according to the SIG's specs than X600, so once there has been another plugfest and ATI jump through the other hoops then presumably there'll be more to come.
What I don't know is whether there are any different requirements for compliance between a component and board level.
For compliance you have a "plug fest" where, AFIAK, you basically try your components/boards out with a bunch of other pcie devices, then run through a bunch of other tests and then run up a load of paperwork based on that test. NVIDIA have decided to do this at the board level, so they have gone through that process with each of their boards to list each one with compliance - by putting X600 through as a "component" they are saying this covers the entire X600 line, as well as board vendor variants for that level of compliance (which, probably purposefully, inadvertently suggests that NVIDIA's vendors would need to go through their own validation for the same level of compliance if they are not using the reference platform from Flextronics [edit] It seems that a bunch of NVIDIA boards are now listed in the components section as well, which would negate that).
X600 is RV380, which is basically ATI's reference platform for PCIe so, unless there were implementation specific issues, its unlikely that any subsequent chip (which is all of them) is going to be less compliant according to the SIG's specs than X600, so once there has been another plugfest and ATI jump through the other hoops then presumably there'll be more to come.
What I don't know is whether there are any different requirements for compliance between a component and board level.
trinibwoy said:
its an internal sales documentation, That a 3rd party decides to leak, and distribute, Unfortunately.
Vortigern_red said:
http://www.pcisig.com/developers/compliance_program/integrators_list/pcie/pcie
Here it is, been posted before, No one seemed at all interested in it however.
karlotta said:
Generally, if something is meant for a private sales team, and then leaked, I find it unfortunate, generally anything private thats made public, I find unfortunate, I dont see how slides like these hurt anyone if they are meant for the internal marketing team. Which they were. The only people who get hurt by these kinds of slides made public are Nvidia, Obviously this kind of marketing would never go public due to various reasons.
So its unfortunate that it got leaked, Its unfortunate that people think that this is some marketing ploy, and even more unfortunate, That people make posts such as yours believing this is actually meant to be public. Its also unfortunate Nvidia doesnt have anymore control over its own internal documents that this keeps happening.
As I have been told, Its for the Nvidia sales team. Which makes perfect sense. People will believe what they want to believe, Which is also unfortunate.
gokickrocks
Regular
ChrisRay said:
it is a marketing ploy, i really dont see how you can see it any other way...target audience may be different, but it is still a marketing ploy
sort of agree with you here...i do not believe it was meant to be an internal document, nor was it to be exposed to the general public, nor was it to be private to the sales team,... i believe it was for manufacturers to see
there is so many things wrong with that statement, i dont know where to begin
Chris - are you aware that X600 is listed in the components section of that list?
And I think Vortigans point is that not being on that list doesn't automatically equate to a failure.
So what you're suggesting basically is that Ati sent x600, only x600 and nothing but the x600 to be tested and just didn't bother sending the other chips/cards cause that would be just too much paperwork and they don't really care about such measley minor OEM checkbox feature.
I dont see whats so wrong with it, since I am talking about the fact that people will probably believe these slides were meant for the public, And yes I still believe its unfortunate that people would believe that,
Internal maybe, Meant for our eyes? Definately not. If this was something Nvidia had all over its website and was spreading across the net with the sole purpose of discrediting ATI hardware, Ya maybe I'd listen. But it's not, So I dont see how it could be considered dirty.
, And Dave, Yes I saw the X600, Originally there were no ATI products tested, or listed as compliant on that list., The question (from my perspective) was he was asking where people would believe ATI wasnt PCI certified or where Nvidia got its "information" regarding this issue. So I simply brought up the list that Nvidia used in its documents (The PCI Sig list)
Is it really important? Probably not. But you and I both know how good it looks on paper.
Mendel said:
The OEM's do their own extensive testing anyway - the likes of Dell will do fairly rigorous testing. This list wasn't even there when ATI were selling their PCIe products and the only reason it has become important is because NVIDIA is busy plugging that their boards are on the list.
A plug fest has just occured last week, and I'm told ATI was there. We'll see if anything more crops up as a result of that.
ChrisRay said:
His question was "Can you point me to a link where there is some proof that ATI have failed the certification tests for PCIe?", not questioning their lack of compliance - Slide 13 boldly makes the claim that "X700 Fails PCI Express Compliance", based on a screenshot of an Inquirer report. Lack of appearing on that list doesn't automatically equate to a failure and yet thats all the evidence that the Inquirer had (and the Inquirer refuted that in a later report).
Anyway, it seems that this isn't the end of it:
http://www.theinquirer.net/?article=18670
http://www.pcisig.com/developers/compliance_program/quality_notice
Dave, There is no reason to use Bold letters, I can just read just fine thank you.
My original statement doesnt change any either, He asked for where people were getting that ATI failed PCIE compatibility, Its clear whether that proof is "incorrect" or not that this is the website that everyone has been referring too. Personally I dont see the problem, Obviously this is the place Nvidia is suggesting that they failed PCIE certification, (According to the slides.)
You however seem to see a problem with this statement. That semantics for you . I'm a little shocked at you though Dave, You should know that this is an internal marketing slide, Yet you are making a mountain out of an ant hill still.
My original statement doesnt change any either, He asked for where people were getting that ATI failed PCIE compatibility, Its clear whether that proof is "incorrect" or not that this is the website that everyone has been referring too. Personally I dont see the problem, Obviously this is the place Nvidia is suggesting that they failed PCIE certification, (According to the slides.)
You however seem to see a problem with this statement. That semantics for you . I'm a little shocked at you though Dave, You should know that this is an internal marketing slide, Yet you are making a mountain out of an ant hill still.
Similar threads
