It's quite clear they speed binned the Geforce3 into a mainstream higher volume part the Ti200 (more chips will hit 175/200 than 200/230 and a lower volumed part (less chips will hit 240/500 than 200/230 and 175/200) Ti500 . Architechturally there isn't any difference in the cards versus Geforce3 . And contrary to popular belief, Nvidia had Geforce4 Ti ready last year. However the timing could have been better. In order to get cards out by the Christmas rush the boardmakers need boards chips by the beginning of October (recent Leadtek interview), so if the chips came back 2 weeks into October or early November they couldn't launch it.
A question regarding NV30 and Doom 3
- Thread starter Sabastian
- Start date
Dunno really. But from my conversation with Tony Tamasi Nvidia's product development manager last year at E3 , it's quite clear in retrospect that he expected the Geforce4 to be their next product, in a few more features and more performance in comparison to Geforce3 (4XS was hinted at, improved pixel shaders, 2 vertex shaders)
My statement on the timing is a guess based on 20/20 hindsight. The information on Geforce4 is not.
My statement on the timing is a guess based on 20/20 hindsight. The information on Geforce4 is not.
Johnny Rotten said:
Yeah if I remember correctly nVidia management bragged(Not that there was anything particularly bad about it.) about their ability to get TSMC to run off the GF3 so that it would run at higher clock rates or something.(Correct me if I am wrong here.) Resulting in the first "refresh" of the GF3. But I remember this because they were saying something along the lines that TSMC would do this only for a preferred customer like nvidia.(nVidia is TSMCs largest customer.) So you are right the GF3Ti series wasn't really a refresh in the sense that others were but it gave nVidia the boost they needed until they were able to role out the GF4 family refresh of the GF3 meanwhile managing large margins on the GF3Ti series "refresh" for relatively the same cost as the GF3. All the while basicly using the same driver set they used with the GF3 while ATi struggled with entirerly new hardware + drivers much in the way nVidia did when they realeased the GF3 in the spring of 2001.
Sabastian
Back in the days of the RIVA 128 and TNT, nVidia stuck to a 12-month architecture cycle. This has since slowed to an 18-month architecure cycle.
Generally, "refresh" parts aren't just different-clocked parts, but actually have modified cores. The GeForce4 was the most significant refresh nVidia has done. Here's how I'd label it:
GeForce SDR/DDR: New core (obviously)GeForce2 GTS/MX: Refresh parts (Two refresh parts for the first time...high-end and low-end).
GeForce2 Ultra/Pro: New clock speeds. No architectural adjustments.
GeForce3: New core
GeForce3 Ti: New clock speeds.
GeForce4 Ti/MX: Refresh parts.
Even back in the days of the RIVA 128ZX and RIVA TNT2, nVidia's refreshes weren't just new parts with different clock speeds (or memory busses, as in the case of the M64 and Vanta), but actually had architectural improvements.
After 18 months, nVidia is most certainly due for another new core.
Generally, "refresh" parts aren't just different-clocked parts, but actually have modified cores. The GeForce4 was the most significant refresh nVidia has done. Here's how I'd label it:
GeForce SDR/DDR: New core (obviously)GeForce2 GTS/MX: Refresh parts (Two refresh parts for the first time...high-end and low-end).
GeForce2 Ultra/Pro: New clock speeds. No architectural adjustments.
GeForce3: New core
GeForce3 Ti: New clock speeds.
GeForce4 Ti/MX: Refresh parts.
Even back in the days of the RIVA 128ZX and RIVA TNT2, nVidia's refreshes weren't just new parts with different clock speeds (or memory busses, as in the case of the M64 and Vanta), but actually had architectural improvements.
After 18 months, nVidia is most certainly due for another new core.
Similar threads
