R350 in .13?
- Thread starter mat
- Start date
No.
tt_22 said:
Simple: margins.
Because 0.13 micron chips are smaller, so in principle they are much cheaper, and therefore attractive for a budget card. Large 0.15 micron chips will be more expensive, therefore not really good candidates for the budget solution (unless you want to lose money). 0.15 can be used to make the high-end stuff which performance-crazy idiots will happily pay through the nose.
You should understand that most of the revenues of these companies comes from the mid- and low-end of the market, not the high-end. So that's where margins are most important.
Conversly, bigger chips have more scope for errors, so using an immature process can result in many more errors meaning the die saving costs may not be offset by yeilds.
nutball said:
I don't really agree with that point of view.
Yes, in time 0.13 will be cheaper than 0.15, but still 0.15 gives much better yields on the chips.
I don't really see why they won't base the RV350 and the R350 on the same core. Maybe i'm wrong about this but i surely hope i'm right.
A 0.13 R350 would surely put Nvidia in shame
I imagine you can get pretty good yields on 0.13u using a less complex ASIC pretty much straight away. TSMC's process should be mature by now.
Both R350 and RV350 are no doubt based on the R300 core, but RV350 will have been migrated to 0.13u, leaned out and packed with as much functionality as they reasonably remove from the PCB.
So will a 0.15u R350.
MuFu.
Both R350 and RV350 are no doubt based on the R300 core, but RV350 will have been migrated to 0.13u, leaned out and packed with as much functionality as they reasonably remove from the PCB.
tt_22 said:
So will a 0.15u R350.
MuFu.
why not - makes sense to me.
test 0.13 on a smaller, less complicated core.
ATI still have a way to go in 0.15, so it'll be cheaper to max that out on an expensive chip, as cheaply as possible, ie, with a high yielding mature process.
Of course as soon as 0.13 is mature and tried & tested it makes no sense to continue a larger process.
test 0.13 on a smaller, less complicated core.
ATI still have a way to go in 0.15, so it'll be cheaper to max that out on an expensive chip, as cheaply as possible, ie, with a high yielding mature process.
Of course as soon as 0.13 is mature and tried & tested it makes no sense to continue a larger process.
Of course, there could also be some other inference from the choices made by sticking with .15u for R350 and .13u for RV350.
Has this been done ever before?Randell said:
Randell said:
I'm pretty sure ATI are updated on TSMC's 0.13 process and know it should be mature by now, ready to make some new sweet R350s.
There is no need to use the 0.15 process if they can use 0.13.
ATI still have the R9700s to improve the 0.15 core on. I don't expect them to drop the R9700 line very soon
I personally think ATI leaked that the R350 is made with 0.15 just to spoil Nvidias NV30 release
They can release a R350 better than the NV30 in every way at the same price, and still sell lots of R9700s.
Anyway i guess Nvidia has lots of homework to do for their next card
There are many thing made on the .13 process..... none of them are 100+ million transistors.... except the GFFX. And look at it's problems! Look at the problems AMD had going to .13... and still has! Seen the yeald figures for Hammer? .13 is not leading edge tech, it's Bleeding edge! ATI is doing the smart thing, as everone here, cept the thread author, is saying.....
martrox said:
0.13 hasn't been very sucessful but it is the future.
Some have made it! Look at intel! The NV30 had its problems but they made it. The corespeed of 500 mhz is very impressive.
The transition to 0.13 will be sooner or later. I am just saying this is the perfect time for ATI to do it. If Nvidia made it, why not ATI
tt_22 said:
NV30 has made it, but look at the troubles it has. It needs that dustbuster to keep the heat down.
Moving to .13u should be easier now though, and if they intend to use the RV350 as a testbed to see how it's going, then it'll give them a good idea of how R400 is going to go.
kid_crisis
Newcomer
DaveBaumann said:
Well, it would make more sense if the 0.13um was actually a different process at a different fab. Then it would be a lot of work to port the R350 over (as opposed to just updating it on the same process).
So the RV350 is being made on UMC's 0.13um process
Randell said:tt_22 said:ATI still have the R9700s to improve the 0.15 core on. I don't expect them to drop the R9700 line very soon
what do you think the R350 is, if not an improved R300 core?
I was thinking improved in case of yields and corespeed. R350 will be an improved R300 core, but with some additional features, like 2 TMUs?, VS3.0?, PS3.0? Even if the R350s are 0.13 they can still improve the R300s, maybe release R9700 ultra with 400 core.
I will be very dissapointed if the R350 is just a R300 with DDR2 memory and higher clockspeed, and I don't think it is.
tt_22 said:Randell said:tt_22 said:ATI still have the R9700s to improve the 0.15 core on. I don't expect them to drop the R9700 line very soon
what do you think the R350 is, if not an improved R300 core?
I was thinking improved in case of yields and corespeed. R350 will be an improved R300 core, but with some additional features, like 2 TMUs?, VS3.0?, PS3.0? Even if the R350s are 0.13 they can still improve the R300s, maybe release R9700 ultra with 400 core.
How would that help? They would just be incurring production costs and polluting their product line with too many similar products. I'm sure part of ATI's reasoning is that they wanted to ensure any FX beating part was not delayed and was prodcued on a reliable process, if core speed is up I'm sure signal integrity is part of the current issues at 0.13 which is less of an issue at lower clock speeds (e.g. value part speeds).
All my speculation of course.