VOGONS


Are GeForce 256 DDR cards that rare?

Topic actions

Reply 280 of 318, by Scali

User metadata
Rank l33t
Rank
l33t
leileilol wrote:

We don't "have to remember" revised history because it surely makes sense they'd release a DX7 card before even DX6.1 came out.

While you are correct in this instance, I do want to point out that there have been instances where the videocard was released before the API it supports.
Examples include the Radeon 9700 vs DX9, GeForce 8800 vs Windows Vista/DX10 and the Geforce 980 vs Windows 10/DX12.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 281 of 318, by leileilol

User metadata
Rank l33t++
Rank
l33t++

In those cases it's more realistic within the time window of a few weeks and they still didn't release before 8.1, 9.0c, etc. on a near-year's worth of pre-api time.

apsosig.png
long live PCem

Reply 282 of 318, by Scali

User metadata
Rank l33t
Rank
l33t
leileilol wrote:

In those cases it's more realistic within the time window of a few weeks and they still didn't release before 8.1, 9.0c, etc. on a near-year's worth of pre-api time.

Well... in the case of the GeForce 980... Introduced in september 2014, DirectX 12 in july 2015.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 284 of 318, by arncht

User metadata
Rank Oldbie
Rank
Oldbie
leileilol wrote:
leileilol wrote:
The Serpent Rider wrote:

You have to remember that NV10 was released in very early 1999 as Quadro. Even Pentium 3 Katmai wasn't released at that time.

Nvidia officially announced it in November 1999....

investigating this "early geforce is quadro its before katmai" bullshit it seems sourced from TechPowerup which assumes the default release date of any video card to be January 1 if it's not specified for its artificially generated articles. Even worse, Wikipedia cites Techpowerup's error as fact and that's okay on their terms because it's a citation!!!1!

We don't "have to remember" revised history because it surely makes sense they'd release a DX7 card before even DX6.1 came out.

Yep, the sdr was available oct-nov, and the the ddr ca from dec in 99 (creative, leadtek), but some vendors released just in q1 of 2000 (asus), not too much before the gts.

My little retro computer world
Overdoze of the demoscene

Reply 285 of 318, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

The 64-bit memory bus GeForce 256 in all its glory: https://www.youtube.com/watch?v=otwyh3JsKyU

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 286 of 318, by 386SX

User metadata
Rank l33t
Rank
l33t

Maybe I have wrong memory of those times but the original Geforce 256 SDR sure was much slower than the DDR version but I wasn't remembering that much compared to those TNT2. Is that possible that with a time correct mobo/cpu/chipset the card would results better?

Reply 288 of 318, by appiah4

User metadata
Rank l33t++
Rank
l33t++
386SX wrote on 2021-09-19, 17:40:

Maybe I have wrong memory of those times but the original Geforce 256 SDR sure was much slower than the DDR version but I wasn't remembering that much compared to those TNT2. Is that possible that with a time correct mobo/cpu/chipset the card would results better?

No, the difference is around 10% I believe. The Geforce 2 GTS (also called the GeForce256 GTS at the time for some reason) was what was almost twice as fast as the SDR..

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 289 of 318, by 386SX

User metadata
Rank l33t
Rank
l33t
devius wrote on 2021-09-19, 18:42:

That's an extra slow version of the GeForce 256 SDR, since it has a 64-bit memory bus instead of 128-bit.

Ok that's why i didn't even remember there was such version but I remembered the original SDR to be a bit better. Is it possible they were trying some cheaper solution MX like, when in fact after the DDR the SDR version soon was left behind? Considering how many TNT2 based solutions there were still around and the soon released GTS..
Anyway i remember different benchmarks even on this forum where the DDR version really was impressively fast to be un "upgraded" version when compared to the SDR 128bit

Reply 290 of 318, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

Is there a good troubleshooting and repair guide for these anywhere? I have the SGI NV990 board and it's had some backside SMDs crunched years ago and halfassed patched, not sure if that was the only prob or whether there was a RAM chip going south and that also needs replacing.

I remember it "sorta working" but seeing the mention of stutter early in the thread, I'm wondering if it was working as built, but I was expecting too much from it at the time I tried it in a system that had the kind of games that ran on a GF4200TI

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 292 of 318, by trixster

User metadata
Rank Newbie
Rank
Newbie

Fits in with the other DDR boards I’ve got

C8F5A829-8CA3-4146-9734-C3D9B163C30A.jpeg
Filename
C8F5A829-8CA3-4146-9734-C3D9B163C30A.jpeg
File size
1.58 MiB
Views
906 views
File license
CC-BY-4.0

I’ve also got this nice ASUS 64MB DDR card too, currently in my PIII 933

2DDE4F71-DABD-49F6-AE39-6225F9EAD3B1.jpeg
Filename
2DDE4F71-DABD-49F6-AE39-6225F9EAD3B1.jpeg
File size
320.52 KiB
Views
906 views
File license
CC-BY-4.0

And here are three SDR models

4C9932AE-1DE6-465B-8B62-AB5B2CE5614D.jpeg
Filename
4C9932AE-1DE6-465B-8B62-AB5B2CE5614D.jpeg
File size
1.78 MiB
Views
906 views
File license
CC-BY-4.0

Reply 293 of 318, by AppleSauce

User metadata
Rank Oldbie
Rank
Oldbie
20211122_180845 (2).jpg
Filename
20211122_180845 (2).jpg
File size
1.06 MiB
Views
899 views
File license
Fair use/fair dealing exception

This is the only Geforce 256 I own.
It's not much but hey its a canopus at least.

Reply 294 of 318, by AppleSauce

User metadata
Rank Oldbie
Rank
Oldbie

Also speaking of TwoFiftySixes , I've got a feeling they're about to get a fair bit more popular since this video came out.

256.PNG
Filename
256.PNG
File size
113.41 KiB
Views
889 views
File license
Public domain

https://www.youtube.com/watch?v=7JEAoKlFvrc

Reply 295 of 318, by 386SX

User metadata
Rank l33t
Rank
l33t

Interesting that power connector I didn't remember some having that already. Does it still stress the AGP bus for the power or it take most from the connector?

Anyway I don't have any DDR version of this card, only the original Creative SDR version. And unfortunately those early fans nowdays make noise.

Reply 296 of 318, by AppleSauce

User metadata
Rank Oldbie
Rank
Oldbie
386SX wrote on 2022-02-12, 12:30:

Interesting that power connector I didn't remember some having that already. Does it still stress the AGP bus for the power or it take most from the connector?

Anyway I don't have any DDR version of this card, only the original Creative SDR version. And unfortunately those early fans nowdays make noise.

I'm not sure unfortunately as I don't have a rig to test it with yet , I'm working on fixing one up though.
I'm assuming it would help somewhat with the load , at least from what I've read?

Reply 297 of 318, by trixster

User metadata
Rank Newbie
Rank
Newbie

That's a very cool card 😀

I thought i'd have a look at 3D Mark 2000 again, with the Geforce256 DD6 64MB that in my PIII 933.

I get 5213 for the default 16bit benchmark. If I use a DDR board in my Athlon 64 X2 3.2ghz agp machine i get 6464 3D Marks, so it's interesting to see how GeForce256 continues to scale beyond a relatively fast PIII from mid 2000.

The other thing I wanted to test was HW T&L vs PIII 933 T&L as I'd read that the CPU quickly beats the GeForce

HW T&L: 5213
Helo Low detail = 106.7
Helo Med detail = 79
Helo High detail = 43.9
Adventure Low detail = 100.5
Adventure Med detail = 64.3
Adventure High detail = 40.3

High poly count 1 light = 11358
High poly count 4 lights = 6506
High poly count 8 lights = 3446

PIII optimized T&L: 4359
Helo Low detail = 79.3
Helo Med detail = 57.6
Helo High detail = 32.5
Adventure Low detail = 99.9
Adventure Med detail = 58.2
Adventure High detail = 35.9

High poly count 1 light = 7039
High poly count 4 lights = 5816
High poly count 8 lights = 4779

It interesting that the PIII is significantly slower than the Geforce in everything other than High poly count 8 lights

----------------

Athlon 64 X2 3.2Ghz, HW T&L: 6464
Helo Low detail = 128.9
Helo Med detail = 94.3
Helo High detail = 45.3
Adventure Low detail = 110.6
Adventure Med detail = 93.7
Adventure High detail = 66.1

High poly count 1 light = 13829
High poly count 4 lights = 6929
High poly count 8 lights = 3684

Athlon 64 X2 3.2Ghz, 3DNow! cpu T&L: 6721
Helo Low detail = 129.8
Helo Med detail = 92.1
Helo High detail = 32.9
Adventure Low detail = 114.3
Adventure Med detail = 103
Adventure High detail = 88.0

High poly count 1 light = 11157
High poly count 4 lights = 11173
High poly count 8 lights = 11168

Last edited by trixster on 2022-02-12, 13:31. Edited 3 times in total.

Reply 298 of 318, by 386SX

User metadata
Rank l33t
Rank
l33t

Cause I think to remember the SDR version on those early mainboards were already quite "heavy" on the bus, that's in an interesting PCB layout. I wonder if other manufacturer built the card without it cause there was still space for the power requirement or close to some limit.

About the above test it'd be interesting to know wich CPU actually surpassed the original Geforce hw T&L speed using sw T&L/MMX/3DNow!/SSE. On the 8 lights test I think to remember old reviews were talking about the impact final results had with that test. I wonder if it's simply too much for the "first GPU" and how much it was increased on the Geforce2 GTS.

Reply 299 of 318, by AppleSauce

User metadata
Rank Oldbie
Rank
Oldbie
trixster wrote on 2022-02-12, 12:33:
That's a very cool card :) […]
Show full quote

That's a very cool card 😀

I thought i'd have a look at 3D Mark 2000 again, with the Geforce256 DD6 64MB that in my PIII 933.

I get 5171 for the default 16bit benchmark. If I use a DDR board in my Athlon 64 X2 3.2ghz agp machine i get 6464 3D Marks, so it's interesting to see how GeForce256 continues to scale beyond a relatively fast PIII from mid 2000.

The other thing I wanted to test was HW T&L vs PIII 933 T&L as I'd read that the CPU quickly beats the GeForce

HW T&L:
Helo Low detail = 103
Helo Med detail = 79
Helo High detail = 43.9
Adventure Low detail = 100.5
Adventure Med detail = 64.3
Adventure High detail = 40.3

High poly count 1 light = 11050
High poly count 4 lights = 6505
High poly count 8 lights = 3446

PIII optimized T&L:
Helo Low detail = 79.3
Helo Med detail = 57.6
Helo High detail = 32.5
Adventure Low detail = 99.9
Adventure Med detail = 58.2
Adventure High detail = 35.9

High poly count 1 light = 7039
High poly count 4 lights = 5816
High poly count 8 lights = 4779

It interesting that the PIII is significantly slower than the Geforce in everything other than High poly count 8 lights

Cheers , It seems like the Hardware TnL is better than most people give it credit , I guess until you hit the 1ghz plus mark.