trodas

But in DOS it works, according to the claims on link that I just added to the post... And on Amiga I do it all the time. Sometimes even halving datacache on 68060 make demo works (previous 68040 have exactly half the datacache, so if demo fail on 060, then that was the next logical thing to do). BTW, there is descripted the asm code to do that in Intel CPU's: http://stackoverflow.com/questions/1108485/disable-l2-l1-caches

Guys, I wonder, if is possible to re-enable L1 and/or L2 caches that are disabled in bios under Windows XP. Specificaly for AMD Athlon / Duron CPU's. These CPU's run deadly slow w/o them (imagine GPU-Z starting 15min - you get the idea what "deadly slow" means) - specifically w/o L1 - and therefore for taking screenshots of the test result I look for means to do it with more "normal" speed and not speed, where just few fast random clicks can cause BSOD, as the input buffer is overflown so easily... (downclocking by nVidia control pannel is the biggest problem) DOS utility that can do it is found there: http://www.vogons.org/viewtopic.php?f=24&t=38613

Well, because CPU-Z crash and I was running Aquamark bench since 13.8. .... see there: These are frames and times of Aquamark 3 running on Vanta LT down to 25/30MHz on Athlon CPU down to 30x9 (yes, 30MHz FSB) with killed caches on MS-6340. (the different writtings that most of the text is my GF) So clearly there is an reason to believe that this will be the slowest Aquamark run ever made. The problems are: - HWbot Aqamark wrapper fail on nVidia Vanta LT (this is the root of all the problems!). - I cannot "just keep trying" of older CPU-Z, because when the run finish, I have only ONE try at this. If it crash, then all I have will be the camera pictures off the screens and GENiEBEN won't create the scores from that, so all will end in vain and to repeat this, I need another like 2 weeks or so of waiting - no old version of CPU-Z will be validated ever, so I will not be able to add validation of the CPU-Z screenshot, untill CPU-Z won't be fixed. - "Why insist on validation for aquamark? Simple screenshot will do." HWbot cannot accept screens anymore, you need to have the validation file. Since you cannot have it, because of the cursed damned bug, the only one way is that GENiEBEN create the scores! And I send him already two older pre-runs to this, with all the screens (score + details) AND all the same screens taken by camera AND all the GPU-Z and GPU-Z windows, just failed to include the screen from the result, as there was no space left AND I was tired from redrawing this too slowly on the screen: ...and no score was build up to this time. So you see, screenshot was not enought, apparently. So there you have the problem... - What makes me angry is, that for some people, the rules don't apply. Above I posted few examples, mainly these two tests did not even have screenshot: http://hwbot.org/submission/627756_ http://hwbot.org/submission/627992_ ...or these scores, where the GPU-Z screens are simply not present: http://hwbot.org/submission/2248070_ http://hwbot.org/submission/2496046_ http://hwbot.org/submission/2349966_ ...and yet, these are valid. When I offer anything besides impossible CPU-Z validation, then it is "not good enought." That is the grief and problem there. lanbonden - And there we go again - that is LIE. If for other people did not GPU-Z work, then they are free to do validation w/o it. See the there fastest WR on ATI Rage 128 PRO above! NONE of them have GPU-Z screen. Should they also - thanks to your "logic", "just use card that does work with GPU-Z"?! WTF... are you even considering what you write?! You see anyone going thru this with their mainboard: http://forum.hwbot.org/showthread.php?t=87051 ...so they are get the "just use mainboard, with what CPU-Z works" BS? Yes, I choose this mainboard, because it can be switched to 30/33MHz depending on CPU used FSB. That is the reason, why this is slowest machine ever. I did not worked my ass on this to be told - just use another machine, problem solved! That is UTTER BS! Thanks, to that I agree wholeheartedly. And that is where we part. Look... to say it calmly one more time: I reported the nVidia Vanta LT bug years ago. Just because it is NOT FIXED BY NOW, is there any problem at all. What else can I do? Okay, yes, as I see it now, I should spendt year on pushing for CPU-Z fix, so I was ready for case, where I'm now. Aquamark is running and the validation won't be accepted, because I will hardly be able to produce a screenshot with the Aquamark result in MS paint (yes, if I just take photo of the details on the screen, that might be done, even it will be about at least 45min work) AND GPU-Z and CPU-Z over it. Let's say I manage put the screenshot in MSpaint and set it on the screen and open GPU-Z. That seems doable. But once we are talking about CPU-Z, there is reall possibility of crash. As I mentioned above: I run CPU-Z on this board there times. Twice it freeze the mobo right away. Once it pretend to run, so I started another instance... and get the freeze again. The older versions of CPU-Z are likely to work, I made (but with SMBus=0 set, hence no Vcore and so on) few screens like that before, but none of these CPU-Z scores get ever validated: http://postimg.org/image/4rrjlrlaz/ http://postimg.org/image/tp5ai81eb/ But what if I remember the settings wrong and the - even older - CPU-Z will crash right away? That is why I asking, if I may replace it with CBId: Central Brain Identifier http://central-brain-identifier.en.softonic.com/ ...just in case that I will wait 2 weeks to see CPU-Z crash on me, taking the score I waited so long with it. Understandable now?

Hmmm, then why it does fail on P4, as above was shown? Is he used wrong AMD OpenCL or what? And I clicked few times into the working windows and on the next batch I get this: ...hope I did not screw up anything. I will stop clicking for sure

havli - Impressive. So a 140.9fps on Game 1 and 105.6fps on Game 2 is possible on default 118/140MHz clocks, when run under Win98 (they are not even SE... ). (the only difference is the 32MB of videoram, mine card have only 16MB... could that hurt the scores so bad?) Therefore your score of 13195 marks are valid and I apologize for not believing that this is possible and I suspected that you must have cheated somehow. My bad. It seems 140.9fps on stock is possible. I still have a little problem with the score, because the score claim to be 13128 marks: http://hwbot.org/submission/2349966_havli_3dmark_99_max_rage_128_pro_13128_marks ...and on the screenshot it is 13195 marks You probably made a typo and lower your score by accident? Second question is about the "random drivers." Is there are reason, why no-one ever use GPU-Z in all the higher, questionable, scores? **** Yes, I still maintain that there are questionable scores, because what you proved was, that 141fps is possible. Not 255fps on stock, as the top score claim! (there is huge difference between 141 and 255fps... but maybe Stermy57 could send you link to the "magic" drivers and we see 255fps? Or anything higher, included Stermy57 162fps is cheat? ... but since between 141fps stock and 162fps overclocked is just a small margin, then I think that Stermy57 score migth be valid too and only one cheater is there - that one, that claims that 255fps is possible on stock ATI Rage 128 PRO, witch I did not believe it is. Or you do? Could someone answer me on this question alone? We now see, that I was wrong and 60fps is not top speed on the card, 141fps is possible... I will recheck that under Win98 when the GPUPI run ends )

Oh, no double precision... That suxx... I was planing oh this move too, but with Socket 775 Celeron

Output file: http://depositfiles.com/files/zmbh5lroj (backup: https://mega.co.nz/#!yUFkWYYb!WGhNcwLN3ukOXCBDkhTJMncWpy-hN0IgPgalk80AemY )

Coool, thanks! ...and this is bauty! One question, tough. How the hell this was cooled?!

GeForce FX 5200 have no image: http://hwbot.org/hardware/videocard/geforce_fx_5200/ That call for fix (nicest image I managed to find, there are plenty of crazy variants, tough) Thanks! :celebration:

No crash here, make sure your PC is stable? Did you checked with OCCT or at least MultiCore Prime95 stress test, starting with FFT size 1792? ( http://postimg.org/image/xg7y7ytrt/ ) http://www.mediafire.com/?5clk6102wkw6ckz (backup: https://mega.co.nz/#!OM0HEDbL!8QgEdEsPgArz9IpqK5iksbEf3LYh1lEuJvjL656b_tI )

Batch 6 finally finished - 48h? What is going on? That way I will be lucky if this run is finished, before are finished the competition... But why such brutal slow-down? I did not touch anythig...

Stermy57 - No offense taken. Just explain me, how stock ATI Rage 128 PRO get to 255fps in 800x600x16, when there is no test or game ever recorded in history that can prove this. Best was max 92fps in Descent... so, how that can be? In same test, 6800 GT does 500fps. So in essence you are telling me, that ATI Rage 128 PRO can on stock do half the vectoring and texturing power of 6800 GT? Come on I view benchmarks as means to evaluate and compare computers speed. Not as tweakable programs, that can be tweaked to get better results w/o any reall hardware improvement. Perhaps that is where we differ? I'm calm, unlike the cause where my best efforts on lowest Aquamark score ever is ruited because of wrapper bugs, this thread does not angers me. It makes me laugh 255fps, ATI Rage 128 PRO - LOL I have enought experience with hardware, starting back in 1990. Amiga 500 since summer 1991, etc. And you want to tell me, what I can or cannot write? Get a grip of reality, mate If I see something wrong, I will say it. Prove me wrong, show me continuous video pan between PC running ATI Rage 128 PRO and then start 3DMark 99 and show me 255fps on stock Rage 128 PRO card and I rest my case. Unless you do it, then I maintain that the card cannot physically do more that about 40fps stock in 800x600x16. So the point is to find a bugged driver that show impossible fps? And you call this benchmarking? Is this why none of the questionable scores show a GPU-Z screen at all? Don't you think that Core 2 Duo 3.4GHz X6800 can overcome the effectivity problems in the driver? But sure you can tell me all about it Or you cannot? ...and? I just took that card as example of same stock clocks, as DrSwizz claims that his ATI Rage 128 PRO have. Like that can change the fact, that no-one can get on stock ATI Rage 128 PRO 255fps in any fullscreen 800x600x16 3D test with contains more that 5 polygons, lol. Well, since your "argument" is that my scores does not count, because I'm not experienced enought, then I took reviews from the most experienced peoples around the world and quess what! No-one cross even 100fps - and that is for 640x480x16 ! :celebration: ...aaah, here we go again. Tualating is more that enought to saturate the Rage card, my friend But once again, I too much stronger CPU into the game and the difference was - non. Therefore you are wrong and I'm right. That is mute point. DrSwiss claimed stock Rage do 255fps. 800x600x16. I fail to find more that 92fps in 640x480x16. Hence it does not look plausible to me, that any other explaination that cheating is possible to make 255fps. (254.8fps, to be precise) ...witch is why you made your score on Win2k :D :D http://img.hwbot.org/u40311/image_id_1111447.jpeg Yea, sure. Nail that 40 to 255fps difference to just a "different driver" that - luckily for you - I cannot obtain. Brilliant! I bet that if you give me PM with link to driver you used, the difference will be under 5fps in the result. Care to try? I cannot wait to see it in action I will gladly apologize for my lack of knowledge how to tweak 3DMark 99 (or, more precisely, how to forcibly disable Vsync, witch could give what? 20, 30% more performace?), yet I still maintain that tweaking benches defeat the purpose of benches. Go on, take the 1st spot, if you show us how you do it. ATM it is only a week, so I wait xxbassplayerxx - I never read about any tweaks yet and it is likely that I stop bench when I know, how things can be fooled... But 100x higher... that is likely an exaggeration, is not it? Yea, but we are not talking about the scores at all. You are mistaken. I talk about the impossibility of ATI Rage 128 PRO on stock clock to drawn 255fps 800x600x16 Game 1 test of 3DMark 99. If this is physically possible, then ATI Rage 128 PRO must have half the 6800 GT speed... 3DMark 01 give +400 points when run under Win2k that with WinXP for ATI Rage 128 PRO. So not 1100-something, but a 1400-something. True. Never tried with Win7... should I? Yes, the 3DMark 99 is likely to run more happy on W2k. Is even possible that using just SP2 or even stock Win2k might yield better results. But within some margins. What my primary argument is, that there are "valid" score that claims 255fps is possible using "stock" clocked 125/143MHz ATI Rage 128 PRO. I might be only amateur at best overclocker, but there is limits of what hardware and and cannot do. 255fps is clearly outside the milits of this old card. I maintain that it will be hard to - after overclocking - reach over 60fps from the card.

What?! Can't you read or what?

GENiEBEN - so, what's the holdup on the two scores, man? I give you EVERYTHING I possibly could. Mr.Scott - What about - if CPU-Z crashes, then you can use alternative program in just this case and just untill CPU-Z become fixed - or the cursed wrapper become fixed? But, for starters, please stop telling me what is not true: "Everybody else manages." That is a monstrous lie. There is NO Aquamark bench made on MS-6340: http://hwbot.org/hardware/motherboard/ms-6340/ You can wonder why. So at first it is not true what you are telling me at all. Second - let me see, what other benches are there, except mine: Christian Ney - 118MHz FSB: http://hwbot.org/submission/2202823_ (CPU-Z used is old, so not comparable with the new one and no Vcore is visible - likely he use the SMBus=0 option to prevent crash - but where you see that guaranted with latest CPU-Z version? Do you really want to tell me, that I lose that superlow (2frames per hour now!) Aquamark bench just because of stupid technicalities that can and WILL get confirmed that they are NOT my fault? Seriously?) Quentj - 850MHz CPU clock: http://hwbot.org/submission/627810_ (again very old CPU-Z v1.40.5 -NO VALIDATION- and with questionable Vcore 2.720V, as the mobo did not have any changable Vcore, so I would strike that value as crazy... but once again, no validation, so... in that case the score is there, but in my case you need EXTRAPERFECT validation? Why that double standards?) Quentj - PiFast score: http://hwbot.org/submission/627652_ (same as above, -NO VALIDATION-) Quentj - SuperPi 1M score: http://hwbot.org/submission/627653_ (same as above, -NO VALIDATION-, now the Vcore is 0.672V for Duron...) Quentj - SuperPi 32M score: http://hwbot.org/submission/627805_ (same as above, -NO VALIDATION-) Quentj - wPrime32M score: http://hwbot.org/submission/627756_'>http://hwbot.org/submission/627756_ (same as above, -NO VALIDATION-, w/o EVEN SCREENSHOT!) Quentj - wPrime1024M score: http://hwbot.org/submission/627992_'>http://hwbot.org/submission/627992_ (same as above, -NO VALIDATION-, w/o EVEN SCREENSHOT!) **************************** So, let me recapitulate it: - there are only THERE people that ever benched this mainboard - one (Christian Ney) submit only highest FSB and that it is - the remaining user Quentj submit scores w/o validation of the CPU-Z at all, sometimes w/o even screenshot required for the wPrime - no one except me did any graphic bench on this board (!) ...yet you telling me, that "everybody else manages" and if I understand it right, you are willing to discard the slowest Aquamark run even made just because: - HWbot wrapper *FAIL* with nVidia Vanta LT, that I reported over 1,5 year ago and it is not fixed by now! - CPU-Z cannot be validated except using latest version now, witch is *NOT* guaranted to NOT crash when run, hence if I try it, it could very likely freeze my machine (run it 3x, once it worked, twice it freezes right away on "Processors"... when it worked, I started CPU-Z again, hoping to do the screen and saving validation - the second copy freeze my computer!) and therefore I will lost the score and have to wait before ever bussy CPU-Z autor will find a time to fix CPU-Z SMBus reading (add exception) for mobo that only 3 people ever benched on HWbot - how high priority will that be?! ... In short, there are plenty of questionable scores (ATI Rage 128 PRO cheaters that claim 255fps on that card on stock - like 1/2 performance of 6800 GT...) and plenty of scores w/o any proof, yet when I do it, you require irefutable proofs, even you know that I cannot provide them. Is that fair? Is this overclocking, or some sort of personal vendeta? What parts of "CPU-Z crashes on MS-6340 mainboard" you did not understand? What can I do to not need to run the Aquamark with such settings as give 30min per frame again? Will you settle with video, showing you repeatedly that on stock settings of the MS-6340 will the CPU-Z crash that I made*AFTER* the bench? Will you allow creating the Aquamark score file from MY informations when Franck confirm that the CPU-Z crash on MS-6340? What proof you require - and yet what little proof have these scores! http://hwbot.org/submission/627756_ http://hwbot.org/submission/627992_

I feeling like that I will be soon cooked alive Temps are closing to 40°C somewhere ( http://novinky.cz/domaci/377625-.html ) and in shadow, my thermometer showing 35°C... ...but my computers have to be keept cool, so I take radical measures

That idea imply different clock settings between all-in-all same bioses... and I very much doubt that they are same. 1) Vanta LT bios is 32 768 bytes long. All other TNT2 bioses are 43 008 bytes and up. 2) Checked differences betwen TnT² Vanta bios v2.05.20.02.00 with Brooktree TV out and the one w/o TV out (third and second from down the table) and the differences are HUGE. There are at end complete blocks that does differ... And there are few small differences everywhere around. I bet that better chance I got to find the downclock limit for GPU... that this.

Oh, such bioses are there: http://www.x86-secret.com/articles/nvbios.htm Sadly it just prove my idea wrong, because for TNT2 cards the X-Bios editor by Ray Adams cannot change the GPU/mem clocks. GF 2 MX400 it can, but not the Vanta LT I want/need...

polandball canada and USA

Guys, for some experimental attempt I'm looking out for VGA bios for some nVidia RIVA TNT2 cards, at best the Vanta LT is preffered: ...do anyone have it and can upload it on for example http://www.mediafire.com or http://dfiles.eu or anywhere (include to reply as attachment)? I need to compare to try lower the GPU clock in the bios Anyone can help? And no... TechPowerUp bioses database did not have anything on TNT2 Vanta cards... They store bioses since FX 6200 cards and not go deeper in history: http://www.techpowerup.com/vgabios/

Go tell my X6800 Extreme, that he have plenty of time, and he even beat your longest run time in the end :-)

I would like to see something more that hearsay, before taking this into account.

This is... somewhat faster that mine

Oh, that is good to know that there is nothign to wory about: ...I was just considered stoping this, because the jump are quite severe. I mean... from around 166 to 176min per loop is suddently jumped between loops 4 and 5 to 846min per loop, witch is roughly 4.8x more But it is true, upon futher inspection, that the slowest run yet had also big jump between batch 4 (18h 36min) and batch 5 (26h 16min), so it is probably okay. I still have to get slower, because I managed only 25h 53min at batch 5, while he have 26h 16min...

I speak too soon. At the end of batch 4 was the test yesterday and today it did not move a single bit...! Still at batch 5, so... That means that there will be much longer loop (just as GENiEBEN say, 15h per loop suddently) and all my "quite regular" loops claims are gone (or more precisely are valid up to the loop 4...) WTF! Question: will be the result accepted or I should strike it down and stop right now? Because it would have a little purpose of wasting time for result, that will be discarded, as there will be sudden and serious jump in time between batch 4 and 5... (and possibly that will repeat somehow during the test) And no, I did not changed anything at all. Two instances of new CPU-Z provide the load on both cores + GPUPI and that it is.

...I quit for year because of HWbot bugs (mainly the cursed Aquamark wrapper...) ... and it looks like that if HWbot team did not get to understand, that CPU-Z crash on some mainboards and therefore it cannot be used as proof, that even my best work on MS-6340 board and Vanta LT go in vain... That make someone consider seriously qutting, because it is impossible to battle that much bugs at once (Aquamark wrapper, CPU-Z, MS-6340 SMBus sensitive to touch, absolutely slow speed, week or two (sometimes it is 2 frames per hour) long benches...).