- Oct 24, 2000
- 29,767
- 32
- 81
ATI held their Annual General Meeting for stock holders yesterday, and while there was some focus on the current OSC investigation into ATI and their CEO KY Ho and a plan to increase the stock options for employee?s, some technical and product details were made available.
Some details learnt at the AGM were passed to me and they appear to back up or conform a number of the details we?ve seen in other stories we?ve reported over the past few months. Here are a few key details:
-There will be a relatively full complement of product refreshes in both the Spring and Fall periods. As expected, Spring will bring R350 and RV350 based products along with a slightly refreshed integrated chipset product (current P4 chipset with faster FSB/RAM support). M10, the mobile equivalent of RV350 (DX9, 4 pipes, 130nm) will be available but its launch will be dependant on customer launches. The fall product will be R400 and possibly RV400 as well (whether this will follow the R400 core, or be a die shrunk R300 is unknown).
-Next Generation of Integrated chipset products will be DX8.1 capable (i.e. Integrated RV250/280 core) and will have dual channel system RAM support, bringing ATI?s chipsets more into the enthusiast realm.
-DirectX9 integrated products are likely for 2004 and can be cost effective on 130nm processes.
It was mentioned that, similar to NVIDIA?s CEO?s recent statements, the intention is to lengthen the product lifecycle such that there will be fewer generational leaps, but larger technological leaps between each generation. Its quite ironic that, seemingly for the first time, ATI are going to hit their much talked of target of 6 months between product cycles and they are already talking about moving off it! It may be the case that they will move to a 9 month cycle, with 18 months between each generational difference.
There are likely to be several drivers for this for an increase in product lifecycle time. First, as seen by the 130nm process introduction, as processes get smaller so too do they become harder to implement, so there will be a greater length of time between implementations. Also, as the PC graphics market become more and more saturated, discrete solution will not be able to sustain the R&D to develop them over such a short period of time. Finally, the big driver for the product refresh cycle is the OEM refresh cycle, and with the PC market in at a general level of saturation it may be the case that OEM cycles are becoming longer.
Also of note was that fact that ATI (and presumably NVIDIA) were warned by TSMC that the path NVIDIA had chosen for NV30 would be best avoided at this stage. Evidently, ATI chose to be conservative and opted for a large design on the 150nm process, which resulted in R300, while NVIDIA chose the aggressive path and opted for it despite TSMC?s advice. ATI estimated that if they sought the 130nm process with low-k dielectrics R300 could have achieved in the order of a 20-25% power reduction and decided that this wasn?t worth the risk.
Source: Beyond3D
Some details learnt at the AGM were passed to me and they appear to back up or conform a number of the details we?ve seen in other stories we?ve reported over the past few months. Here are a few key details:
-There will be a relatively full complement of product refreshes in both the Spring and Fall periods. As expected, Spring will bring R350 and RV350 based products along with a slightly refreshed integrated chipset product (current P4 chipset with faster FSB/RAM support). M10, the mobile equivalent of RV350 (DX9, 4 pipes, 130nm) will be available but its launch will be dependant on customer launches. The fall product will be R400 and possibly RV400 as well (whether this will follow the R400 core, or be a die shrunk R300 is unknown).
-Next Generation of Integrated chipset products will be DX8.1 capable (i.e. Integrated RV250/280 core) and will have dual channel system RAM support, bringing ATI?s chipsets more into the enthusiast realm.
-DirectX9 integrated products are likely for 2004 and can be cost effective on 130nm processes.
It was mentioned that, similar to NVIDIA?s CEO?s recent statements, the intention is to lengthen the product lifecycle such that there will be fewer generational leaps, but larger technological leaps between each generation. Its quite ironic that, seemingly for the first time, ATI are going to hit their much talked of target of 6 months between product cycles and they are already talking about moving off it! It may be the case that they will move to a 9 month cycle, with 18 months between each generational difference.
There are likely to be several drivers for this for an increase in product lifecycle time. First, as seen by the 130nm process introduction, as processes get smaller so too do they become harder to implement, so there will be a greater length of time between implementations. Also, as the PC graphics market become more and more saturated, discrete solution will not be able to sustain the R&D to develop them over such a short period of time. Finally, the big driver for the product refresh cycle is the OEM refresh cycle, and with the PC market in at a general level of saturation it may be the case that OEM cycles are becoming longer.
Also of note was that fact that ATI (and presumably NVIDIA) were warned by TSMC that the path NVIDIA had chosen for NV30 would be best avoided at this stage. Evidently, ATI chose to be conservative and opted for a large design on the 150nm process, which resulted in R300, while NVIDIA chose the aggressive path and opted for it despite TSMC?s advice. ATI estimated that if they sought the 130nm process with low-k dielectrics R300 could have achieved in the order of a 20-25% power reduction and decided that this wasn?t worth the risk.
Source: Beyond3D
Facebook
Twitter