DrSwizz
-
Posts
111 -
Joined
-
Last visited
Content Type
Profiles
Forums
Events
Blogs
Posts posted by DrSwizz
-
-
Yes, it certainly is one amazing card!
When I first started overclocking it (after I had modded it) I started Ati-Tool and used the "find max core"-feature (over whatever it is called). The clock speed started at 50Mhz (I guess that is the default for 2D-mode). Ati-Tool slowly, but surely kept increasing the gpu core speed. When it reached 400MHz I thought "this can't be right, Ati-Tool must be broken". Eventually windows crashed somewhere above 540Mhz.
-
You have to try over and over
, for me it took 2 weeks :D. As I have 3 Bench System it was OKI tried 7-8 times with more or less the same result.
I suppose I could try harder to find a different version of the drivers or I could disable direct draw acceleration entirely. And I have second Rage Pro Turbo card that I also could try. I suppose I more recent version of directx wouldn't hurt either.
-
Unfortunately increasing the timeout value didn't help much. :-(
The explosion seemed to progress a little further, but windows still killed the driver. I guess the upper limit for the timeout must be something lower than 65536.
I thinking of giving up on this card and aquamark now.
-
Yes I have searched it too, cause it's the driver who is crashing at about 5050 fps of the bench, it's beaucause the card is not as powerful, she doesn't support the explosion
. That why I have underclocked it, doing so she is as unpowerful that the explosion is not calculated, so she can reach the end of the bench Hmm... So the big explosion at the end of the benchmark is removed when the card is being clocked lower?
When I look at windows/aquamark behaves when the big explosion is being rendered it seems that the driver is doing just fine, but since the driver is so busy with the explosion Windows thinks that the driver has crashed and therefore kills/restarts the driver.
Is is possible to somehow increase the timeout before windows kills/restarts the driver?
Edit:
I did some googling and found this:
http://technet.microsoft.com/sv-se/library/cc976062(en-us).aspx
I have set the value to 65536 on my rig now, I think that should be enough ;-)
I wonder if changing this value is enough, though. Could there be other timers in the kernel that could cause problems?
-
@ DrSwizz, Impossible to run 3DMark01, and Aquamark always crashing after 16 hours for me that why I have downclocked it, the drivers were : for WindowsXP :
http://support.amd.com/us/gpudownload/Pages/legacy-xp.aspx
Select the 2nd one ( Rage Pro ).
Yes, it is too bad 3dmark01 requires soo much video memory. :-/
Thank for the link. I am currently using those drivers too.
I asked about it because I was figuring that it might be the drivers that was causing the crashes and I have looked in vain for a slightly newer/older versions of the drivers.
-
Nice!
What drivers are you using for that card?
I am currently benching a Rage Pro Turbo myself, but so far aquamark has always crashed after 5-6 hours :-(
-
Now I have apparently contributed 100 points to my nonexisting team.
You might want to check so that there isn't a bug that causes my points to become attributed to some random team (not that it really matters to me personally, but it would not be fair to all the other teams).
-
After a lot of search i finally find good mobo for 'edo' stage.
It is a FIC VA-503+ based on VIA MVP3 chipset.
It has 4 simm connectors and capable up to 124 MHz fsb with 1:1 to memory.
But the main reason was: cpu-z can correctly show memory type (EDO) and frequency!
My 60-ns sticks have maximum working freq. between 100 and 112 MHz (next jumper setting).
At 112 windows cannot boot.
Lucky Star 5MVP3 might be also a goot one, mine have dead keydoard port.
Have you tried using software to OC the system bus?
The PLL on my MVP3 boards has 103MHz setting that I can only access using software. :-)
Some notes on WinNT 4.0 sp6: surprisingly it has no mouse emulation through num keypad (and i've lost my com mouse lol).
And mose strange: printscreen doesn't work for some reason, i have to search and use ancient screenshot utility.
For me CPU-Z caused the mouse to lock up in NT4 (I used a serial mouse). So I tried using a PS/2 mouse instead, but then both the mouse and the keyboard locked up, very annoying.
Printscreen did work for me though, but windows paint didn't work very well without a mouse. :-/
@speddy411: Nice score.
-
100MHz! Not bad at all. :-)
-
How do you do that, CPU-Z doesn't show me the mem frequency on my PII even PIII
That is odd. I have not had such problems with CPU-Z except with really old hardware such as the first Pentium and K6 CPUs.
The only advice I can think of giving you is to check your chipset drivers, they might be the cause of the problem.
Also ask Antinomy for help, he seems to know allot about the inner workings of CPU-Z and hardware in general. :-)
Good job
Thanks!
-
yeah great score. unfortuntely i also have never managed to see the memory clock in cpu-z
Use a K6-2 CPU: CPU-Z is able to get the bus speed and multiplier readings from those. :-)
-
-
Ok, problem with ST6E solved: actually cpu-z doesn't work properly witn fsb: ram: pci = 3:3:1,
and shows corrently with 4:4:1.
I tried with amd k6 setup (aladdin 5) ant p mmx (430TX), it also doesn't show memory type.
May be dimm edo witn 440BX will be recognized properly?
Waiting for custom cpu-z wersion?
Or how we can participate in stage 3?
CPU-Z probably relies on an EEPROM to be available on the DIMMs.
I had a bit of problem with this myself: I found a really old DIMM that lacked the eeprom. The only way I found to indicate what kind of memory it was (EDO or SDRAM) was that the stick didn't work with SDRAM-only boards.
-
Today i tried my Abit ST6E (i815), 175 fsb stable,
but unfortunately cpu-z can't show even 1:1 mem divider...
Time to blow a dust out of TUSL and cross fingers?
By the way is it possible to make a cpu-z submission with EDO ram at all or not?
Seems to be it isn't support entirely, who had success with it?
This is the only submission I have made with EDO so far:
http://valid.canardpc.com/show_oc.php?id=1277286
as you can see CPU-Z doesn't recognize the RAM (or much else). It might work better with newer boards though as that particular board & CPU is ancient.
-
;63773']i will try it but i have the slightest impression that i've ran that program from a floppy in DOS mode a few time ago to find out the manufacturer of a mainboard but i may be wrong.
i have another soldered am386dx-40 exactly like this one but i cant find any kind of jumper configuration for the mainboard that has it. one of those programs that identify bios strings say its a Seritech Enterprises mainboard' date=' i've googled it and found only 4 bios strings from mainboards of this manufacturer, 3 of the bios strings were connected to a mainboard and the other bios string (equal to mine) was from an unknown mainboard, now that's bad luck. the mainboard i'm talking about seems almost like a new out-of-the-box mainboard and has the capability of supporting a 387 math coprocessor installed together with the 386 chip but without jumper configurations i can't swap cache chips for some better ones and i can't overclock the cpu
. i have a i386dx-33 chip too but i don't have a mainboard for it .[/quote']Too bad you are having trouble finding relevant documentation.
I where you I would look at the jumper settings for the other boards from the same manufacturer and look for similarly placed jumpers on your board. You might be able to figure out & guess enough settings to get the board running. :-)
;63774']started pifast now' date=' i'm feeling that in a few days (lol) this thread will receive an update
[/quote']Good luck!
-
It is good to see a 386 alive and kicking!
I can highly recommend a program called "syschk": I have an old copy of it on floppy disk that I have been using from time to time.
I did a google search and to my surprise there actually was a homepage for that application: http://www.syschk.com/
It provides plenty of other system information too and seems to behave well running under windows.
Stange: I only posted my reply one time, but yet it ended up being posted twice.
Something must be wonky with the hwbot server.
-
It is good to see a 386 alive and kicking!
I can highly recommend a program called "syschk": I have an old copy of it on floppy disk that I have been using from time to time.
I did a google search and to my surprise there actually was a homepage for that application: http://www.syschk.com/
It provides plenty of other system information too and seems to behave well running under windows.
-
For some unknown reason I now got the "contribute 50 points to your team" achievement without ever having been member of a team.
Also when I look at the page associated with this achievement it is all blank:
http://hwbot.org/community/achievement/1061_contributing_to_the_team
-
I don't see any problems in your submission. You mean it doesn't show the default frequency? Well, Linux isn't able too - it can make you think it does
Take a socket 7, change the multi and Linux will get confused too.
Yes, in Linux the stock frequency (60MHz) as well as the current (overclocked: 66Mhz) frequency was shown, I was a little surprised when I first saw that.And since you seemed to be looking for input on how CPU-Z behaved on older hardware I figured I mention it.
You mean that Linux performs some kind of trick to figure out that OCed my CPU? ;-)
One more thing: Yesterday evening I uploaded that text-output you wanted: I only edited my last message on the competion-thread so I didn't show up as new message, so you might have missed it.
-
Yes, it's informative because a very simple reason - the 486 especially Intel's did have the CPUID feature but they showed the registers values only on reset - that's why BIOS can figure out the CPU but software is not so precise.
Very smart decision.
I hope the popularity of this family will increase.
P.S. can you tell, whether CPU-Z is able to recognize any of your 486? I see you have ones from Intel, AMD and Cyrix. If yes, a picture of CPU-Z would be nice
CPU-Z recognition of the original Pentium CPU (the 800nm, 5V CPU) needs to be improved;
It shows up as a 500nm CPU and there is no indication of which frequency the CPU was made for whereas /proc/cpuinfo in Linux shows the correct value.
-
DrSwizz, you uploaded two validation files. I asked to the "TXT report" it's on the last tab, the "About" one.
Sorry; I guess didn't read your request carefully enough.
I'll upload that text-output soon.
Edit: I have uploaded the textfiles you asked for here:
http://rapidshare.com/files/406399770/mvp3g2_memsettings.tar.gz
:-)
-
As I said - two CPU-Z reports in txt with only the memory speed (synch and asynch) being changed would be very nice.
And please, name the files correspondent. Or type a comment which one is it.
Reports from more people is always better than from one - to be sure.
I have uploaded two CPU-Z files here with ~100MHz bus frequency and different divider/ratio for the memory here:
http://rapidshare.com/files/406339016/mvp3_memsettings.tar.gz
They are from version 1.54 though, does that matter?
-
Make a txt report using CPU-Z on FSB100 when memory is at 100MHz and 66MHz - you know, the Mem=AGP mode. If possible, change only this setting, so it doesn't harden the decryption. And attach here for example.
If you still have need for CPU-Z output from an MVP3-board I am doing some benchmarking on such board right now. :-)
-
@Turrican: I am impressed that you had the patience to wait for the benchmark to finnish. 50 days is a long time! Good work!
DrSwizz - Radeon 7000 SDR ~ Radeon VE SDR @ 540/176MHz - 83 marks 3Dmark 2003
in Result Discussions
Posted
Thanks
Actually I am not that much familiar with everest (I have only used the memory/cache benchmark). I'll have to check what "sepcs" is. :-)