Unknown Soldier
Veteran
OpenGL guy said:
Yes .. and if you get attacked.. attack back where they are the strongest.. and that would be Desktop Graphics Cards. The R300 is a sweet card hey!!
US
Yes .. and if you get attacked.. attack back where they are the strongest.. and that would be Desktop Graphics Cards. The R300 is a sweet card hey!!
US
artisan7 said:SanGreal said:Ascended Saiyan said:
-----------------------------------
If they already had final silicon it would be a tremendous waste of money & time to scrap it in order to make major revisions to it. From what I understand, its not exactly just changing a few numbers in a computer. More likely they would've released the card then refreshed it with improvements later.
DOnt you think Guys that BIG companies ,
like INtel , NVidia /AMD or others are ,
are a little more smarter than the tiny brains from
ATi Gamers here ?
<snipped more ranting>
All I said was I don't think a company would throw away a final silicon because it would be a huge waste of resources?? I don't see how that required a page long rant about how stupid everyone is?
Given the '90-100 days Tape-out to market' numbers are being quoted I think thats a clearsign they are talking about first tape out.
http://www.hardocp.com/
http://www.hardocp.com/
Unknown Soldier
Veteran
Ye i read that too .. hopefully that includes the 256bit bus
Damn .. can't wait to see wot this babe can do
US
Damn .. can't wait to see wot this babe can do
US
OpenGL guy
Veteran
Dave, I am not convinced. Four of their chips took three revs. That's at least 3 months in the fab, more likely 4 to 5 if we assume they retaped out immediately when the chips got back. The numbers don't add up...DaveBaumann said:
Unknown Soldier
Veteran
OpenGL guy said:
OpenGL .. well if they do Tapeout(read it says first tapeout) it might just be to see if there are any flaws .. and if there are .. to fix it and if there ain't to continue. I suppose that's wher the A01, A02 numbers come from . .although i could be wrong.
US
OpenGL guy
Veteran
Whenever you tape out, you assume it's going to be your last... it's very expensive to create the masks required to make chips. If the chips come back with bugs you can't deal with in software, then you have to fix the major problems and do another tape out. This all takes time...Unknown Soldier said:
Doomtrooper
Veteran
http://www.eedesign.com/story/OEG20020612S0051
This was a summary of a refresh or the Geforce 4...
63 million transistors, 78% logic
50% of RTL from GF3 modified
400k lines C , 800k lines Verilog
Schedule targets: 9 months (!!) to tapeout, 100 days from tapeout to ramp
40 to 70 engineers in implementation alone
$160 million in tools, $40 million of which on emulation
Total cost 100-160 million dollars to .15 to .15 .... First silicon had 19 reported functional problems, only seven of which required repair. All seven were corrected in the metal, allowing the team to meet its target of approximately 9 months to tape out and 100 days between tape-out and production ramp.
This was a summary of a refresh or the Geforce 4...
63 million transistors, 78% logic
50% of RTL from GF3 modified
400k lines C , 800k lines Verilog
Schedule targets: 9 months (!!) to tapeout, 100 days from tapeout to ramp
40 to 70 engineers in implementation alone
$160 million in tools, $40 million of which on emulation
Total cost 100-160 million dollars to .15 to .15 .... First silicon had 19 reported functional problems, only seven of which required repair. All seven were corrected in the metal, allowing the team to meet its target of approximately 9 months to tape out and 100 days between tape-out and production ramp.
This is CEO speak - he's not going to the the analyst "Well, this is a new design and new designs usually have more bugs." He can say typically or on average we take 90-100 days from tape out to market, because they do - no need to worry your investors more than you have to.
OpenGL guy
Veteran
Exactly. My point is that I believe he is talking about the final tape out. Again, I am speculating just like anyone else, but looking at their track record, it sure looks like the previous numbers are referring to final tape out.DaveBaumann said:
I guess we'll know more in a couple months
Edit: I realized I wasn't being more clear. I think the CEO meant they haven't taped out anything. If they had, I believe they would say so in order to instill confidence in investors. Again, I believe the time to market numbers for the other parts are from final tape out to full production.
I was under the impression that the typical rule of thumb for a 'repspin' was in the region of 4 (if you're lucky) to 6 weeks - whick is about 28-42 days; would you then have another 58+ days for production ramp up after that?
artisan7 said:
I think you'll find there is far less 'fanatic' speculation here - try reading some of the fan sites to see what real fanatic speculation is like.
Oh, and please, learn how to type on the forum correctly - your posts are a mess and very annoying to read!
nooneyouknow said:
Actually, I don't think that is the case either. I merely suggested that he might have used the term "tape out" in a more broad sense than we expect and thus saying something along the lines of "we are close to a full go in the NV30". But again, I don't think he was.
nooneyouknow said:
I think you're right on this one too. The fact is, however, that they have talked about the NV30 as a fall part and during Spring talked about the NV30 as being the first DX9 part to hit the market. They also didn't correct the mistake that people (e.g. investors) thought that the NV30 did tape out in June. This is what I called "a bit fishy".
Time will tell, time will tell...
Unknown Soldier said:
There were 2 reasons I beleive the Gf3 was slower in some cases -
drivers & the fact that it doesn't like SDR systems. Aceshardware has some upgrade articles which shows just throwing a Gf3/4 in an SDRAM system it is slower than a Gf2Pro say. The Gf3 is now faster in those games it was losing to the Gf2U in on modern systems with enough power and memory badnwidth .
The LMA and nfiniteFX are integral parts of the Gf3 architecture soI fail to see how they were put into the Gf3 to 'speed it up over the gf2'
In respect of the NV30, true DX9 games will be a way off, we have no knowledge of how either the R300 or the NV30 will actually perform using DX9 applications.
As for Cg will speed it up over it competitiors - dont even go there
Just one more thought (then I'll shut up!):
Maybe, just maybe, the NV30 isn't really delayed according to nVidia's own internal timetable.
nVidia might have decided that they wouldn't ship the NV30 before DX 9 was ready. They also have to wait to the DDR-II is ready just like they had to make sure that TSMC was ready to roll with their 0.13 process (which is crucial to the NV30 design).
They might just have said: "hell, we cannot ship this chip before November anyway, so let's continue with the design phase a bit longer and use some more time to test in simulation".
It's a risky business because they have to be fairly certain that the first silicon back is very close to production quality - e.g. no more than one fairly easy remake of the metal mask is needed.
I don't believe, however, that there was any delay due to "hey, we need to add a 256 bit bus also!". The major design cannot be changed, but some minor tweaking after the test simulation might have taken place during the last month or so.
Maybe, just maybe, the NV30 isn't really delayed according to nVidia's own internal timetable.
nVidia might have decided that they wouldn't ship the NV30 before DX 9 was ready. They also have to wait to the DDR-II is ready just like they had to make sure that TSMC was ready to roll with their 0.13 process (which is crucial to the NV30 design).
They might just have said: "hell, we cannot ship this chip before November anyway, so let's continue with the design phase a bit longer and use some more time to test in simulation".
It's a risky business because they have to be fairly certain that the first silicon back is very close to production quality - e.g. no more than one fairly easy remake of the metal mask is needed.
I don't believe, however, that there was any delay due to "hey, we need to add a 256 bit bus also!". The major design cannot be changed, but some minor tweaking after the test simulation might have taken place during the last month or so.
artisan7 said:
No, big companies are no smarter than everything else. That's why some of them succeed and others go out of business. If all companies were smart, then none of them would ever fail. Show me this overclocked GF4 and show me the benchmarks where it is faster than R300. Your post is just utter nonsense.
