VOGONS


First post, by Wes1262

User metadata
Rank Member
Rank
Member

This site claims that the Geforce 2 Ti (Released together with the Geforce 3 Ti, so way past the Geforce 2 life) had a shrunken NV15

https://www.anandtech.com/show/873/2

However I remember from back in the day reading that Nvidia had a bunch of unsold NV15s, which they rebranded as the entry level card to the GF3Ti line.

Which to me makes more sense than shrinking a core that at the time was already old.

Who is right? Unfortunately looking under the heatsink won't reveal the truth, as the nVidia cores have a lid.

Thanks.

Reply 1 of 5, by rmay635703

User metadata
Rank Oldbie
Rank
Oldbie

Honestly the GeForce 3 didn’t have much of any lifespan, it was manufactured maybe a year?

Pretty much 3 was an intermediate mid to high end GeForce “2” until GeForce 4 came out.

Reply 2 of 5, by Wes1262

User metadata
Rank Member
Rank
Member

Everything was moving very fast back then, yes. There are entire lines of cards that lasted only 6 months, like HD 2xxxx. Literally didn't have time to hit the market that the 3870 was already released.
That being said, there is a big difference between 2 3 and 4..... they were moving fast and improving fast, a lot!

Reply 3 of 5, by auron

User metadata
Rank Oldbie
Rank
Oldbie

at least nvidia's product page makes no mention of it, treating the ti in line with the other geforce 2 series cards: https://web.archive.org/web/20011204201753/ht … AGE=geforce2pro however, in their "product overview" PDF they also list a maximum of 32 mb for the GTS which is obviously wrong.

from NV15 images i can find, originally each product had dedicated writing on the chip, but later on things apparently got more unified. these ultra and ti examples look basically identical:
https://www.vgamuseum.info/index.php/cpu/item … -geforce2-ultra
https://www.vgamuseum.info/index.php/cards/it … dia-geforce2-ti

so judging by that, maybe it's the other way around and 150nm parts also wound up on later ultra cards, even before ti was launched? presumably, 150nm would offer better yields in terms of hitting the 250mhz clock. another question is if they used a lower voltage for their assumed 150nm process chips - that could give a hint, for instance comparing a GTS with a ti.

edit: okay, some chips with ti written on them do actually exist: https://thandor.net/object/593

Last edited by auron on 2024-07-23, 01:11. Edited 1 time in total.

Reply 4 of 5, by Wes1262

User metadata
Rank Member
Rank
Member
auron wrote on 2024-07-22, 21:27:
at least nvidia's product page makes no mention of it, treating the ti in line with the other geforce 2 series cards: https://we […]
Show full quote

at least nvidia's product page makes no mention of it, treating the ti in line with the other geforce 2 series cards: https://web.archive.org/web/20011204201753/ht … AGE=geforce2pro however, in their "product overview" they also list a maximum of 32 mb for the GTS which is obviously wrong.

from NV15 images i can find, originally each product had dedicated writing on the chip, but later on things apparently got more unified. these ultra and ti examples look basically identical:
https://www.vgamuseum.info/index.php/cpu/item … -geforce2-ultra
https://www.vgamuseum.info/index.php/cards/it … dia-geforce2-ti

so judging by that, maybe it's the other way around and 150nm parts also wound up on later ultra cards, even before ti was launched? presumably, 150nm would offer better yields in terms of hitting the 250mhz clock. another question is if they used a lower voltage for their assumed 150nm process chips - that could give a hint, for instance comparing a GTS with a ti.

Hard to say without looking under the nvidia lid thingy. Maybe with a thermal camera it would be possible to estimate whether the die is the same or smaller without breaking it open.