RS600/RS690 to be based on M26
- Thread starter Arty
- Start date
Kaleidoscope.
However, I'm wondering about this. The distinction between RV410 and RV370 is some minor shader capabilities and a bunch of pipelines - RS480/400 is half and RV370, so is RS600 going to be half RV410, in which case it'll actually be closer RV370 as it is, but with PS2.b! RV410 also has a large number of VS's for its class all of which will never make it across to an integrated chipset, but will some? Its about time integrated boards got some hardware vertex shaders.
However, I'm wondering about this. The distinction between RV410 and RV370 is some minor shader capabilities and a bunch of pipelines - RS480/400 is half and RV370, so is RS600 going to be half RV410, in which case it'll actually be closer RV370 as it is, but with PS2.b! RV410 also has a large number of VS's for its class all of which will never make it across to an integrated chipset, but will some? Its about time integrated boards got some hardware vertex shaders.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2222geo said:
Dave post is very interesting. Me goes searching for Intel/Nvidia's solution slated for the same time frame ...
One of the things that really stuck me (and that, in itself, is notable as to how hard they were hitting the point, as usually chipset stuff is borderline boring to me) about the last conference call, is the combination of both pride and alarm on the chipset side. Pride in that they feel they are getting very good market reception --better even than they were expecting, is the tone I was getting.
Alarm in that it sounded like they discovered there just is no pricing power in that market --the price you can sell at "is what it is", and the best features you can stuff in for that price, then great --but if you put in more than that and your costs go up. . .well, too bad, the price points will not move for you, no matter how much better the product is than its competition.
They talked several times (without getting into detail), and with a great deal of intensity, about their plans to get the margins up. . .and based on the above, it sounded to me like only two options. . .controlling costs better, and taking features out. Maybe 90nm helps on the former as it matures, but I'm wondering if we are going to start seeing them cut back on capabilities on some of these integrated compared to what they were originally planning.
Alarm in that it sounded like they discovered there just is no pricing power in that market --the price you can sell at "is what it is", and the best features you can stuff in for that price, then great --but if you put in more than that and your costs go up. . .well, too bad, the price points will not move for you, no matter how much better the product is than its competition.
They talked several times (without getting into detail), and with a great deal of intensity, about their plans to get the margins up. . .and based on the above, it sounded to me like only two options. . .controlling costs better, and taking features out. Maybe 90nm helps on the former as it matures, but I'm wondering if we are going to start seeing them cut back on capabilities on some of these integrated compared to what they were originally planning.
I would think that the link to R520 indicates that RV515 will eventually replace RV410 as the integrated graphics core. Otherwise ATI will see itself fighting the SM3.0 checkmark battle all over again, at a time when the number of games using it will likely be counted on more than one hand. I find it altogether strange that they aren't going straight to RV515 considering these chipsets are slated for 2006 and that NVDA will be competing with its own SM3.0 integrated core very soon.
No, I'd put money on it being Kaleidoscope. There has been frequent references in the past for the "low end" to offer more functionality/features than the higher end 3D stuff get in relation, especially on the video side. this would be a move into getting very high quality video support directly into an integrated chipset which will be very useful for Media Center and HTPC systems, alongside the standard desktop wins.
(IMO Anandtech got what "Kaleidoscope" was wrong, but only because that was attached to the "product" they thought it was)
(IMO Anandtech got what "Kaleidoscope" was wrong, but only because that was attached to the "product" they thought it was)
I consider this as disappointment.
when going from RV350 to RS400, they halved the number of pipelines. ( 4 -> 2 )
so if they are going to do this again, they will have 4-pipeline shader model 2 chip,
ie. chip that's quite identical to RV350 chip I've been having in my laptop for more than 2 years,
and would mean that integrated is lagging the mainstream mobile market by about 3 years.
.. and shader model 3 chip only in some distant future.
integrated chip based on RV515 would be quite a lot nicer but propably it would be too big..
though there is the dotted line from R520 to RS600, lets hope it means it will have some improvements from the R5xx-series..
when going from RV350 to RS400, they halved the number of pipelines. ( 4 -> 2 )
so if they are going to do this again, they will have 4-pipeline shader model 2 chip,
ie. chip that's quite identical to RV350 chip I've been having in my laptop for more than 2 years,
and would mean that integrated is lagging the mainstream mobile market by about 3 years.
.. and shader model 3 chip only in some distant future.
integrated chip based on RV515 would be quite a lot nicer but propably it would be too big..
though there is the dotted line from R520 to RS600, lets hope it means it will have some improvements from the R5xx-series..
Similar threads
