• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

9500NP is now harder to mod, I believe

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Maggot

Registered
Joined
Jan 12, 2002
Location
Goldens Bridge
The rules of the game have definitely changed... for the worse.

You guys had a thread here 2 weeks ago about "hobbling" of the 9500NP, but the single criterion cited was the "L-shaped RAM" as indicating that this is a 9700 PCB. I believe that's all over now.

http://www.rage3d.com/board/editpost.php?s=&action=editpost&postid=1331958914

above thread got lots of hits, but no comments and no credibility- AFAIK- but I'm not looking for applause: this news sucks. I'll leave it here like a dead fish at your doorstep and then I'll let it drop.
The post I wrote went like this:


It used to be that you just needed to get a 9500NP with the 9700 PCB -meaning that the ram chips had to be in an L shape, not straight. That was enough to indicate you'd have a fair shot at opening up the 4 pipes just waiting , and the 256 bit width of ram too.

Well, I had a sapphire 9500NP with "the 9700 PCB" (& I still have it- it's a great paperweight). It came a month ago. It had L-shaped ram (Infineon 3.3 ns) and it took the soft mod with ease: fillrates were superb and it OC'd so well that I decided to watercool the core and sink the ram chips. I wanted to convert it to a hardmod prior to permanent attachment of a waterblock. I believe that was when I ruined it: the soldering job required more delicacy than I could provide. Nothing I tried to save the card worked- & I tried everything to save that beauty, but even the POST ascii looked like Braille cyrillic.

So- still not content with my failure and the lessons it offered- I decided to try again. I got another 9500NP. It arrived 2 weeks ago with Hy 3.6 ns ram- again in L-shape.

But... that wasn't all that had changed: now there are different IC's on the flip side of the card!
Opposite the auxilliary power jack, there'd been 2 identical 14-pin ICs: ISL6522CB/P301ACQ but the newer 9500NP had one 14-pin (ISL-same/P240BEH) *and* an 8-pin: 3037A/CS233 and there were 2 new components opposite (and flanking) the GPU site:
old had 106C/P425S, new had 10/16V.

The old card had stencilled on to it V2A, and P/N ended with 073520 and the new had V39, P/N suffix 067378.

The new one took the softmod and opened its pipes- but it also had artifacts of a peculiar type: 2 vertical pairs of parallel dashes, one on each side of the monitor just like these, but vertical:

============================

They looked stuck-on to me: a deliberate spoiler that just served to blemish an otherwise beautiful rendering job.

I believe the changes in the PCB layout have been exclusively for the purpose of deliberately spoiling the mod.

So now the suitability of
a 9500NP to take the mod is no longer just a crapshoot of getting a board that was binned yet still works the full pipes: the pipes work just fine, but the "artifact" -which disappears when the mod is reversed BTW- is a protection ATI has developed.
 
Damn.
Wait tho. The vertical lines, are they just on the sides of the screen? How big are they? Cause if they are small you could adjust the monitor to strech the image so they dont show... And what effect does adjusting the resoulotion have on them?

Maybe its a "feature" in the card's bios... In which case, and hacked bios would fix it...
 
Well this is certainly bad news I didnt want. I just ordered a 9500np with the intention of opening the extra pipes with the softmod....:eek: :eek: :eek: :eek:
I can understand (and im not angry) ATI pulling the plug because they are in the business of selling performance graphics cards, not handing out free ones at half price....still, I hope I was early enough to catch the bus:(
 
Then again, what about 9500 Pro?
Did they change that? 9700NP/P??!??!?!

Maggot,
I assume it was the same company?
Try changing the resouloutions and see what happens. Try lots of settings, Ansiotropic, Anti-Aliasing.... so on and so forth.
 
Is running this card with artifacts bad for it? I get checkerboarding, but it's not severe. If I leave the other 4 enabled, will it damage the card, or just bother me? Also, I hear that checkerboarding can be remedied by disabling the hierarchal(sp) Z unit . . . What exactly is this contraption, what does it do, and what happens if I disable it?

Also, before anyone starts to jump to conclusions, I think we need to check to see if everyone with this same cards has this problem. It COULD be a legitimate artifact, one which ATI has no control over . . . Also, why would ATI bother to design a card that inhibits this procedure, if they already axed the L-shaped memory? Seems like a waste of effort and R&D money . . .?

Thanks

Z
 
It sounds more like you just got one with a broken pipeline. ATI abandoned the 9500 production already so it is extremely unlikely that they did anything to the card.

One thing people dont seem to understand about the 9500 mod... IT DOESNT ALWAYS WORK!!! I've had 3 sapphire 9500 cards all with infineon 3.3 in the L. They all had different extremes of artifacting, but none the less had it.
 
Yea, I figure that while maybe a great percentage of the people that post have success, the numbers are probably more in favor of unsuccessful. It's not supposed to work ats a 9700, so it's not surprising that they don't. Honestly, I doubt that ATI was throwing 9700s into the 9500 bin. I think some of the 9500s just turned out without defects and went unchecked.

Rocko- did you run your cards with defects or did you return them?

Z
 
RMA'd the first 2, this 3rd one I'm still using because its limited. The only place I have noticed artifacts currently is in the nature demo of 3dmark 03 and in splinter cell on the oil derrick mission. In both cases the 'transparent' water is completely checkerboarded. All other games I play and tests I run are fine though.

And I believe ATI was throwing 9700s into the 9500s bin... but I believe they were ALL BROKEN 9700s. Theres obviously something wrong with the cards that they did not meet QA for 9700 status and were downgraded. IMO even people who claim to have perfect 'successful' mods are mistaken and will eventually find something that does not run correctly. I thought this 3rd card was a perfect card until I saw those 2 very screwed up examples.
 
Are you trying to get rid of the checkerboarding, or are you just trying to live with it? The checkerboarding isn't really bad on mine, and it only shows up in certain places. I haven't tried to game with it yet, though.

And by saying that they weren't putting 9700s in the 9500 bin, I just meant that I don't believe they knowingly put WORKING 9700s in the 9500 bin. That would have just been stupid.

Z
 
Guys, guys, you're missing my point!

The most poignant fact is my observation of changes in the hardware of the card: the different IC's I described in successive samples of the same card (Sapphire red 9500NP, 128) : different ram silicon (not a new observation) and different chips (IC's I've not seen described elsewhere).

I RMA'd the card so I cannot experiment further, but it'd surprise me if this deliberate hobbling could be resolved by a simple change in settings.

I was anxious to solicit posts from other recent buyers to tell if they have seen the chip config I had: it was a shock to me that No One but me had reported the change in the hardware or the odd new "artifact".

You are correct to observe that the production run was reported to be small, and perhaps complete, so the glory days of 9500NP are done- yet will join the fond memories of Cel300A, 440BX, P4-1.8, (and all the AMD ones I stupidly ignored)!
 
I'm just living with it now because as I said it is limited to those 2 cases. Since the replay value of splinter cell is pretty much 0 after beating it twice on hard, and since I dont care what a benchmark looks like, I see no reason to try to fix it. When future games show similar artifacts I'll just can the card and buy something else.
 
So are you thinking, like me, that the checkerboarding doesn't do damage to the card, that it's only a sign of a manufacturing "glitch"? I don't really mind the checkerboarding, but I'm sure as he** not going to wreck my card if it will damage it to run with the checkerboarding.

I guess we'll never know for sure if it will do damage.

Z
 
Rocko[DPC] said:
RMA'd the first 2, this 3rd one I'm still using because its limited. The only place I have noticed artifacts currently is in the nature demo of 3dmark 03 and in splinter cell on the oil derrick mission. In both cases the 'transparent' water is completely checkerboarded. All other games I play and tests I run are fine though.

And I believe ATI was throwing 9700s into the 9500s bin... but I believe they were ALL BROKEN 9700s. Theres obviously something wrong with the cards that they did not meet QA for 9700 status and were downgraded.
that's flat-out WRONG. Have you ever hard of cpu-speed-binning?. goes something like this - cpus are tested to find stable speeds. some are of course worse than others, so they get designated as lower cpus, while the better ones end up as the to-end chips.

The problem with that, is that yeilds are often so good, that the manufacturers (amd for instance) dont have enough "bad" cpus to meet the demands of the lower speed chips. Therfor, they take these better performing cpus and call them xp1700's for instance, even though they can run very much faster.

Well, the same thing happens with gpu's. You can bet your right arm that ATI couldnt find enough "bad" gpus to meet the demands of the 9500's, so those fully working gpus ended up being used.
IMO even people who claim to have perfect 'successful' mods are mistaken and will eventually find something that does not run correctly. I thought this 3rd card was a perfect card until I saw those 2 very screwed up examples.
All i can say about that, is my softmodded r9500 has run everything thrown at it without a single artifact since the day i brought it (about 3-4months ago i think).
 
james.miller said:

that's flat-out WRONG.

Do you work for ATI? I said that is what I believe. You can believe what you want.

james.miller said:

All i can say about that, is my softmodded r9500 has run everything thrown at it without a single artifact since the day i brought it (about 3-4months ago i think).

Good for you. 'I believe' that although you do not see anything wrong with the card, there is in fact a reason why it was downgraded.

I understand how the situation works with AMD. There is a high demand for the low end chips and they must meet that demand. I do not see a high demand for 9500 nps. If youre not going to unlock the other 4 pipelines, the 9500 pro is a much better value card.

This is only my opinion. As such you cannot call it flat-out WRONG. Youre welcome to have your own opinion.
 
Rocko[DPC] said:


Do you work for ATI? I said that is what I believe. You can believe what you want.
No, i dont.
Good for you. 'I believe' that although you do not see anything wrong with the card, there is in fact a reason why it was downgraded.
Umm, yeah. Like i said, to meet demand.
I understand how the situation works with AMD. There is a high demand for the low end chips and they must meet that demand. I do not see a high demand for 9500 nps. If youre not going to unlock the other 4 pipelines, the 9500 pro is a much better value card.
Look harder then. ALOT of people have brought 9500non-pro for the sole purpose of trying the mod. There are people on these forums who have brought 3 or more 9500np's to get a succsessfull card. there are also people who buy them because they are cheap, good performers ect.ect.
This is only my opinion. As such you cannot call it flat-out WRONG. Youre welcome to have your own opinion.
And so are you. Difference is, my opinion is based on FACT.

Tell me this, how many threads have there been about people's modded cards suddenly getting checkerboards or a mod-related problem? hmmm. How 'bout none.
 
Last edited:
james.miller said:


Tell me this, how many threads have there been about people's modded cards suddenly getting checkerboards or a mod-related problem? hmmm. How 'bout none.

Umm...

How about this thread?
 
what about it? he tried to alter the hardware and destroyed the board.Now you KNOW thats not what i meant:rolleyes: I meant a softmod-related problem.

as for its replacement - the softmod never fully worked in the first place. try reading the thread again, will you?
 
You asked for 'a mod-related problem'. This thread is a mod-related problem.

Why do I have to read the thread again?

Since this is getting off topic I wont reply anymore. If you wanna continue babbling theres a PM button.
 
jesus chris:rolleyes:
how many threads have there been about people's modded cards suddenly getting checkerboards or a mod-related problem?

you DONT know what suddenly means? as in, suddenly doing something it never did before? oh well........not my problem.
 
James is right. The guy messed up his card while adding a waterblock, NOT because of a softmod.
 
Back