Latest HWiNFO32 crash on Asus TXP4-X mobo

[trodas]

Hello, Martin   I salute the good HWiNFO autor and I have to report that HWiNFO badly crash on the end of HW detection on Asus TXP4-X maniboard. Yes, it is old Socket 5/7 mainboard, but it have some interesting features (mainly the ability to run at 7.14MHz FSB) and I recapped it completely in 2008 for friend: http://trodas.wz.cz/index.php?act=ST&f=16&t=462

Now I get it from him and started to overclock it: P90 at 75x1.5 give 112.5MHz and run SuperPi 32MB well:

http://hwbot.org/submission/2962892_

However since I cannot find a program, that report reliably the FSB (CPU-Z is a joke, it says 55x2: http://valid.x86.fr/ia4gwh
...i tried your great HWiNFO. However HWiNFO twice crashed to BSOD,
powering off HDD each time and second time even erasing the BIOS settings,
so only the debug is produced.

Included.

The only oddity there is, that the bios is flashed from Asus TX97-XE, so the mobo is reported as Asus TX97-XE, while it is Asus TXP4-X in reality:


But that is, of course, not any matter for this bugreport, as this is a result of my tweaks   So, the mobo is stable, even overclocked, yet HWiNFO crash (and CPU-Z is reporting crazy things about the FSB and not reporting rams at all, but Aida64 does report the rams correctly) - what can I do to help overcome this crash?

Run some beta? Disable something...? Let me know, Martin, please.

 

[Martin]

It's odd, since the DBG doesn't show any sign of a crash and completes successful. Maybe the BSOD was deferred after HWiNFO finished the scan..
My advice is to try to disable Disk scan to see if that helps maybe.

 

[trodas]

It is worth the try. The HDD is too old... but for better ones I need moded bios that allow for >32G HDDs, so... Will do it and report back.

 

[trodas]

Sadly, no luck. I disabled the option "Disable Drive Scan" and it still crash and power off the HDD. But I'm ready for another attempt - what could I disable next, please?

 

[Martin]

Well, I'm not sure as the DBG doesn't reveal where exactly the crash happens. Try to change several settings in the Safety section.

 

[VictorVG]

Sorry, but you have it in any museum of antiquities found computer? Payment judging by your photographs are not OK - in chains VRM swelled filter capacitors (typical fault in this series of motherboards), in addition to it you connect incompatible devices Zoo. Okay, S3 Trio 32/64 working on it, but for the HL-DT-STDVD-RAM GH22NP20 need a newer chipset, at least class VIA KT400/600/880 with AMD Athlon XP/MP 2200+ for DVD-RAM recorder LG (in nee Gold Star) to work with i430TX somehow not designed - this generation chipsets are not able to provide the necessary for its operation bandwidth in principle, but less CPU Pentium !!! 733/AMD Athlon 700 did not fit, and to put them in the S5/7, you will have them "extra" bite the legs. They really do not run, but it will rise.

Hence, I like the design engineer computer only one conclusion - you absolutely do not understand the technology and that you Cray XT7, that tin is all one, just to label "computer" was.

The conclusion is not in your favor, and about the "bug report" - the morgue for meaningless piece of iron - in a landfill for repairs it will rise a lot more expensive than buying a suit on the board at DDR3-1333 Intel P45 / X48 with Core 2 Duo E8600 together with the selection of memory and NVIDIA Quadro K6000.

Look up and place MBProbe v1.30 Beta 9 and it will then operate without interruption. This utility is compatible with WinXP, but you'll find it also in the museum.

P.S.

But if without jokes and talk seriously, then your design is initially not operable as a consequence of damage to the system board time you select beyond the tolerance modes of two nodes, and the use of it is not applicable for this equipment units - three. In general, suitable for souvenirs, no more. Yes, and the modification of the firmware (BIOS), you could damage it as code that can cause crashes. And even if you put on video boad SOJ to increase its memory up to 1MB good of it will not. You're wasting my time, Martin and offering to "fix" does not exist error HWiNFO32 caused by your own miscalculations when creating this model sample.

It's enough. The motherboard is not serviceable and its repair is not economically profitable - the new card will cost ten times cheaper than the details for this.

 

[trodas]

Martin -

Well, I'm not sure as the DBG doesn't reveal where exactly the crash happens. Try to change several settings in the Safety section.

Okay, will do disable more. Just when current SuperPi 32M run ends, the bloody K5 need about 1h per loop, 24 loops... go figure. And that is overclocked at 90MHz Slow



VictorVG - you did not understand the point, mate. The point is not to restore or make better the old hardware for profit, the point is to have fun, have a good Socket 7 maniboard and gain some knowledge in the end You know that there are competitions that:

- require submit SuperPi 32M runs that took over 12h?
- require submiting lowest Intel and AMD CPU clocks?

No? Then you did not understand anything I was trying to do. This board is a golden thing, because it can be forced to use as FSB clock the reference /2 CLOCK by it's PLL chip. The reference crystal being 14.3MHz, so giving 7.14MHz FSB.

Now I waiting for SuperPi 32M to finish on K5 and then I try to solder the crystal to the PS/2 controller to make the keyboard and mouse works in the 7.14MHz FSB clock. Hopefully I suceed.

Next step then will be 32MHz bus.

If I understand that right, then it is possible to lock on the PCI bus to 32MHz regardless of the FSB clock. That would allow me to overclock into 83MHz FSB w/o problems (unreliable IDE controller) on the PCI bus...!

Selecting BSEL = 0 could be pretty cool mod and I cannot understand, why Asus did not do it already. When using PLL chip that can do PCI lock - it is almost a crime not to utilize that...!

judging by your photographs are not OK - in chains VRM swelled filter capacitors (typical fault in this series of motherboards)

If you actually bother to look futher down, you see, that I already recapped all the faulty caps by quality caps. That was first step. Also you might noticed, that I made sucesfull SuperPi 32M runs on this mainboard, witch lasted sometimes more that 19h:
http://hwbot.org/submission/2965531_

Can even your computer last SuperPi 32M calculation over 19h sucesfully?

incompatible devices ... the HL-DT-STDVD-RAM GH22NP20 need a newer chipset

Blah! It works w/o any troubles.

in a landfill for repairs it will rise a lot more expensive than buying a suit on the board at DDR3-1333 Intel P45 / X48 with Core 2 Duo E8600 together with the selection of memory and NVIDIA Quadro K6000

But can P45 / X48 with Core 2 Duo E8600 run at 10.7MHz as this beauty does?

http://valid.x86.fr/9b88hu

And can this P45 / X48 mobo run AMD CPU at 10.7MHz? And can that mobo produce 12h+ SuperPi 32M runs?

...if you have to answer no to all these questions, then you know why I did not want to replace any mobo for THIS particular one. Make me offer and see, if I care

Look up and place MBProbe v1.30 Beta 9 and it will then operate without interruption.

Finally I hear something constructive there. I will try it, once the SPi run ends (loop 17 ATM, 7 loops to go, so about 7h to go):
http://www.majorgeeks.com/mg/getmirror/mbprobe,2.html

your design is initially not operable

My tests beg to differ

Yes, and the modification of the firmware (BIOS), you could damage it as code that can cause crashes.

LOL ... you are even qualified to be around cumputers, much less to operate them? I just wonder... :huh:  No, seriously. Do you even know, what BIOS it and what it do? Because if you do, you cannot write such a comment. The TX97-XE differ only very slightly and I can put the original bios back, no problemo. I don't want argue with you, but are you willing to even do as little, as to look on WHOLE page and not just first few pics?
And - are you willing to apologize publicaly, when the original bios did not fix HWiNFO crash?

You're wasting my time

I'm so sorry about it. But... where did I pressure you to reply? And yes, I want HWiNFO not fail on this board. Is that a crime now?

The motherboard is not serviceable and its repair is not economically profitable - the new card will cost ten times cheaper than the details for this

I already repaired it. Long time ago. It proven to be stable by SuperPi 32M tests. Repeatedly. What more...  okay, so you are going to send me a new mainboard to replace this one? So, what about Gigagbyte P45 UD3P ...? I will send you my address by PM, where you can send the board, and I did not even need that you include the E8600 or rams.
Then I drop this thread and bugreport... or do you are just talking nonsenses, or you are willing to put the money where your mounth is?

 

[VictorVG]

With the selection of RAM will help - at most there was a problem in the MSI P45D3 Platinum (MS7513) when ran into 2 * Corsair CMX2GX3M1A1333C9 + 2 * Micron MT16JTF25664AZ (inherited from the burnt-out car on the P43 with the C2D E7300) - car crashed and fell into a panic the kernel until Corsair modules replaced by 2 * Samsung M378B5673GB0-CH9 as they work exclusively in a pair and can be put only one pair, but with no guarantee that they will be compatible with your motherboard - complaints on these modules in the forum of the Corsair more than enough. For chipsets P43/P45/X48 DDR3 modules need to put 1 GB or 2 GB (up to 8 Gb summary) single-sided double-sided two / four bank-modules with memory chips organization 4 or 8 bits, the modern 16-bit and not run out of DDR2 (up to 16 Gb DDR2-800 or 8 Gb DDR2-1067, can accidentally came across a couple of great ) at a frequency of 1112 MHz (tested instrument itself this fact surprised) have launched and allowed to set the frequency of 800 MHz RAM Kingston KVR800D2N6/2G. However, pre-tip - at FSB 1333MHz P45 northbridge gets very hot, so it is it is further cooled. For example, in the ASUS P5QL/EPU (P43), P5Q SE (P45) with a standard radiator at a frequency of 800 MHz, it is heated to 40 - 45, 1066 with the already up to 55 - 60, 1333 and for him not to touch - already it is heated by the radiator 100, if not more, and saves only the outer air cooling fan to the CPU. On the MSI MS7513 with its powerful cooling system of no - 1333 North Bridge on a calibrated XK (Chromel-Copel) thermocouple is heated to just 38 on the lid of the chip!

When choosing DDR2 RAM can focus on the frequency that puts the chipset without setting:

FSB = 800 - RAM = 800
FSB = 1067 - RAM = 1067
FSB = 1333 - RAM = 1112

For DDR3 he puts the ratio of 1: 1. If you put the modules with different SPD, it will have more to raise the frequency FSB (I Current Memory Clock: 676.2 MHz (2: 1 ratio)) that will lift and dissipate CPU heat, but I put the Thermaltake Big Typhoon with ArcticCooler F12 PWM2 (regular long- thrown out because of the noise) and even in the heat of the CPU 43 on the street I do not heated above +48 at full load when the counting and databases and compilers, and more that are considered to be heavy at the same time clock. Well, it is clear in the case several 120 mm fans to cool the air flow through the machine work differently it will be difficult - total heat in the housing station at about a 500 W ...

With regard to the said board - nice and solidly made. Since it was no problem. There is no cooling in the economy, while Aluminum radiators, but the thermal conductivity of aluminum is not inferior to copper - he only lower heat capacity, but with the amount of heat that is necessary to dispel the north bridge of the cooling system lacks.

 

[trodas]

Asus TXP4-X MBprobe (latest 1.31 b9) show 0 sensors, but at least it show CPU voltage:



...and does not crash, not even with the AMD K5 PR75

 

[VictorVG]

Please, install driver giveio.sys otherwise birthday grandmother shows. I remember that since those times when hex editor tinkered with its translation into Russian.