View Full Version : softquadro 4 is out!!!
http://nvworld.ru/docs/sq4e.html
|
|
GregHess
11-18-2002, 02:51 AM
I'm getting mixed reports of app instability in max with these. Be forewarned. Have not heard anything about any other apps. (Max's ogl is pretty crappy off the bat). But I wouldn't consider testing these out on any production level machine. (Try it out on a machine you can waste a few hours fixing if something goes wrong)
Also note that in some of their tests the "softquadro" results are SIGNIFICANTLY slower then the non softquadro cards...
[I've just seen a tiny sampling of cases...you could have absolutely no trouble and get a massive performance leap. Always good to be forewarned though. Better to say "Greg your full of $%#(*, then to sit there crying cause you won't finish your work in time cause your OS is hosed.]
I guess those extra million or so transistors on the quadro cards must do something after all. :)
dvornik
11-18-2002, 03:14 AM
Woo-hooo! Thanks.
[edit] I agree about max 5 - it's OGL is just slow, Quadro or not.
MaDSheeP
11-18-2002, 12:14 PM
damn... that is something that would be very cool to have.. with the potential of hosing your system... =(
i think i will just save up for a 750XGL
oh.. does anyone know the gaming performance with the Quadro cards?
GregHess
11-18-2002, 12:19 PM
I doubt it will hose your system. Just should always be careful anytime your using any soft of bypasser, or "hack".
Always better to er on the side of safety (Backing up all your current work for example). Then to sit there trying out how to get max reauthorized cause you had to reinstall it.
I'm just having some flashbacks to earlier softquadro hacks (rivatuner), where a few people on the max forum had a really #$@% time with them...before elsa had discreet shut down all conversation about softquadro on the discreet forums.
MaDSheeP
11-18-2002, 12:26 PM
heh, actually.. my last computer got struck by lightening.. and i built a new one about 2 months ago.. installed xp pro... loaded everything up, and then installed Max4... and the bizzare thing is, it never asked me to authorize...
CStudio asked me and i did e-mail discreet and mentioned that it never asked me to authorize it.. they didn't mention it in their reply e-mail with my cstudio key...
i was using my old hard drive that did have an authorized copy of Max4... but that is now being used as a slave.. heh, just thought it was funny...
oh, and.. i have killed my windows installations too many times trying out things =(
i would LOVE to give the softquadro a try, but things like that scare me now
Originally posted by GregHess
Also note that in some of their tests the "softquadro" results are SIGNIFICANTLY slower then the non softquadro cards...
Hi Greg
Looking at these benchmarks,
SoftQuadro4 vs. geforce4 on P4 2.26 with Ti4200
First (http://nvworld.ru/pics/sq4/image013.gif)
Second (http://nvworld.ru/pics/sq4/image017.gif)
Third (http://nvworld.ru/pics/sq4/image021.gif)
Fourth (http://nvworld.ru/pics/sq4/image025.gif)
(this one either geforce4 beats soft4 or more likely I think labelling is wrong from reading the following paragraph in the review.)
Fifth (http://nvworld.ru/pics/sq4/image029.gif)
I don't see any reason not to try the hack.
By Wednedsay, I'll have the same setup. Could you please link to the tests where they show a decrease in performance ?
I intend to enable this hack as soon as the card arrives and since I don't play games that much, it doesn't matter. But my key question is suppose something goes wrong or doesn't go as expected, what's the safe way to reinstall the original drivers ?
(Note that I would prefer not reinstalling my OS again, so a safe way, however convoluted is welcome. I'm using w2k sp3)
Thanks
Gyan
GregHess
11-18-2002, 03:43 PM
Hey Gyan,
The charts have been changed. (Just looking at it now again). I'll have to go home and check my cached pages to point out which one was originally showing the geforce 4 beating the softquadro. Your probably correct about it just being mislabeled.
I'm currently searching for the other sites with "softquadro results" looks like the articles aren't posted anymore. I'll keep looking.
"But my key question is suppose something goes wrong or doesn't go as expected, what's the safe way to reinstall the original drivers ?"
Usually going into win2k Safe mode and uninstalling the hardware (device manager, uninstall geforce X).
Sometimes you can also reset the res of the default 2k install if it goes out of wack by checking the display properties in safe mode and moving the resolution slider around and hitting apply. (It will sometimes popup a window that says..."Set resolution to Defaults?"
dvornik
11-18-2002, 06:25 PM
You may remember Kondakov's tests for Unwinder
http://www.digit-life.com/articles2/profcards/profcards-09-2002-svp.html
and earlier:
http://www.digit-life.com/articles/profcards/profcards-04-2002-update1.html
Can anyone run some SpecAPC tests on it? I guess it would make a difference in max 4 especially with maxtreme, but not in max5.
http://www.specbench.org/gpc/apc.static/max42info.html
GregHess
11-18-2002, 07:06 PM
Ah thast the one dvornik, Thanks!
Did you ever try the updated maxtreme for max5? Before I left the discreet forums some people were talking about a version that worked with 5.
dvornik
11-18-2002, 08:28 PM
I could make 4.13 work, sort of, with hardly any performance gains and some visual artifacts. They also mentioned 4.30 but I couldn't find it. The rest didn't work for me. The reason I mention it is that maxtreme seems to signifcantly improve SpecAPC results in max 4.2. I think Specviewperf itself is a bit misleading for max users.
loop29
11-18-2002, 08:37 PM
This is not giving you too much problems, the main thing is the nvstrap driver which is installed and loaded before OS kernel, it maps deviceID of GF4 to Quadro4, after that OS determines your GF4 as Quadro4. After installing nvstrap your OS will find new graphics hardware and you must install driver which you must patch with antiprotection scripts and softquadro4 script, donīt forget to unpack driver with the batch file before you patch, otherwise it will not patch correctly, I was beta testing the SQ4 on WinXP and Win2000 and it doesnīt gave me any problems with install or uninstall. If you wanna get rid of the nvstrap driver just uninstall the driver through Rivatuner GUI and you will be ready to reinstall GF4. The performance of Gf4 with SQ4 in accelerated lines and two sided lightning is slower than original Quadro4, but as mentioned in the article it is not clear if this comes from feature blocking in the driver or if it is hardware difference. Anyway, with SQ4 applied you can install maxtreme driver for 3ds max which is not possible with GF4, I only tested Specapc for 3dsmax 4.26. System: AMD Athlon XP 1800+, 512 MB DDRAM CL2 266 MHz, Raid 0 80 Gbyte, Geforce 4 Ti 4400 (XGL 750) -->Graphic mean: 6.7 , CPU mean: 3.81, Overall geometric mean: 5.5
regards
dvornik
11-19-2002, 02:01 AM
Interesting. My setup is similar, only with a real 750 XGL, 1900+, 512 DDR, 40.72, Max 4.25, Maxtreme 4.00.21, 1280x1024. and I get
Graphic mean 5.87
CPU mean 2.74
Overall 4.5
But their 4200 on P4 2.26 beat my setup on many Specviewperf tests without SQ, I guess my machine is quite a mess now.
What did you get without Maxtreme or SQ4 (I'm not sure I can wait for that damn table to move)?
[edit] Looking at my old results after fresh install I was getting up to 4.38 CPU and 5.11 overall (29.80). But still...
loop29
11-19-2002, 07:46 AM
Yes, I benchmarked this P4 system, this beats the crap out of AthlonXP systems. But remember that this system had RDRAM 1066 and FSB 533, no wonder why an Athlon system canīt be on par with this monster. You tried maxtrem 4.00.25, Iīm not completely sure but I think this maxtreme version was optimized for Quadro 4.
regards
GregHess
11-19-2002, 01:52 PM
Ya P4's with rambus used to be king of the viewports, while the athlons held the crown of rendering.
Now it appears the two switched...with the fastest viewport system being a 2800+ XP with an nforce2 board, while the fastest rendering rig is an HT 3.06 P4.
All depends on app as well.
paultheplumber
11-19-2002, 10:16 PM
Originally posted by GregHess
Now it appears the two switched...with the fastest viewport system being a 2800+ XP with an nforce2 board, while the fastest rendering rig is an HT 3.06 P4. In the Softquadro article concerning the P4's increadible % performance jump compared to the AMD:
"It is almost assured, that such benchmarking results are consequence of the perfectly made driver optimization for SSE2."
If this is the case, wouldn't a P4 system run faster viewports on a system with a Quadro/SoftQuadro graphics card? It'd be iteresting to see nForce 2 vrs GraniteBay vrs a few older boards just to see what the price/performance ratio is...
So many benchmarks.. so little time... :hmm:
GregHess
11-19-2002, 10:32 PM
All the benchmarks thus far have shown the 2800+ To be signifcantly faster then viritually everything in the viewports.
http://www.tech-report.com/reviews/2002q4/pentium4-3.06/index.x?pg=10
Even with a 800 megahertz advantage, not only does the 3.06 HT processor lose, it loses badly in some of the tests.
And this is with specviewperf, which I believe is actually SSE biased.
(Specviewperf tends to recreate a bit better the average viewport tests/manipulations we do on a daily basis...then say something like ut 2003, or quake3)
Of course this is just viewport manipulation studies..
Here's some secondary and teritary data collaboration....(remember were just looking at viewport stuff here, not rendering and what not, I've already stated that the 3.06 is a superior rendering proc)
http://www.xbitlabs.com/cpu/pentium4-3066/index2.html#12
http://www.anandtech.com/cpu/showdoc.html?i=1746&p=19
loop29
11-19-2002, 10:42 PM
OMG, Athlon XP 2800+ king of the hill in viewport performance ? Thatīs what I wanna see, incredible. The pushed FSB of 333 MHz seems to really help the Athlon XP getting itīs job done. But even with this higher FSB it will not reach the bandwidth of a P4 system with 1066 RDRAM which is a proof for being P4 the better rendering system....
regards
GregHess
11-19-2002, 11:05 PM
Ya its really odd that both of them kinda of flipflopped.. While the p4's were increasing their rendering powers, the athlons slipped right past them in the viewport.
It would have been interesting had AMD integrated SSE2 into their Barton core...oh baby would that have pissed off intel.
CGTalk Moderation
01-13-2006, 09:00 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.
vBulletin v3.0.5, Copyright ©2000-2013, Jelsoft Enterprises Ltd.