TPU's Nostalgic Hardware Club

[dorsetknob]

Trying to see if this HP Touchsmart IQ770 will run, does that count as old?

Main Fault they had was the MMX style Graphics card and its Soldergate Problem
If your lucky it was replaced/sorted before you acquired it

 

[Toothless]

Main Fault they had was the MMX style Graphics card and its Soldergate Problem
If your lucky it was replaced/sorted before you acquired it

Well it posts into the BIOS just fine, currently trying to see if it'll do the boot into windows thing with an SSD.

 

[lexluthermiester]

THEY WERE ALSO BEASTS FOR OVERCLOCKING ( not unusual to get a 100% Overclock OR MORE with additional cooling)

Right. And that 333 Celeron will got straight to 100mhz FSB if you knock the multi down to 4.5, no voltage tweaks or extra cooling needed, effectively becoming a Celeron 300A. If you kept the multi at 5 you could get 500mhz out of it, but most of them needed a voltage bump and a bigger fan on the cooler.

Here's an interesting video, examining performance degredation over successive drivers using WIndows 98 and a series of late 90s and early 2000s nVIDIA cards. Shocking to see how bad the later drivers are.

There's actually a logical explanation for this(as it happened to the Radeons too). As functionality and additional code were added to the driver set the processing over-head increased. This is still happening to a lesser degree with current gen cards and drivers. Thus, if your driver is working stable, don't update it(if it's not broken don't fix it).

 

[delshay]

That Asrock motherboard is the stuff of dreams, sad that its not really possible nowadays to implement such thing. Asrock are pretty new when they release that, its a pretty bold move by them when no other big names making something similar.

My Asrock 939 with capacitor change to all polymer. This is the first time i'v got a FX60 dual core this high.

Another thing to note here is the image quality. This shows up not just by using the snipping tool, but also on the IGP & GFX card.
Polymer capacitors has changed this. Everything is much more defined & detailed.

 

[Trekkie4]

Right. And that 333 Celeron will got straight to 100mhz FSB if you knock the multi down to 4.5, no voltage tweaks or extra cooling needed, effectively becoming a Celeron 300A. If you kept the multi at 5 you could get 500mhz out of it, but most of them needed a voltage bump and a bigger fan on the cooler.


There's actually a logical explanation for this(as it happened to the Radeons too). As functionality and additional code were added to the driver set the processing over-head increased. This is still happening to a lesser degree with current gen cards and drivers. Thus, if your driver is working stable, don't update it(if it's not broken don't fix it).

Ah yes - the good ol' days...

 

[lexluthermiester]

Ah yes - the good ol' days...

Nice! Noisy little buggers though. I always preferred mounting an 80mm or 90mm fan to the top of the stock heat sink and calling it good. But these always worked well to OC with.

 

[Toothless]

IQ770 works, though it heats up my room better than both of my towers running WCG. There might be an issue when you open the door to my room and you can FEEL where the heat is coming from. Good thing I found my MX-2 paste.

 

[agent_x007]

My Asrock 939 with capacitor change to all polymer. This is the first time i'v got a FX60 dual core this high.

And here's my ASRock 939Dual-VSTA in action : Windows 10 x64 on Socket 939 motherboard

Even LAN works (after Windows Update searched online for driver, because Smartphone + USB Internet Sharing = Internet)
Oh, you may or may not belive me, but this screenshot was done on R9 390X (card isn't visible and can't be installed because of "Code 35*" error with PCI-e bridge [*System Firmware doesn't have info needed to make it work] :/).
Here's valid : https://valid.x86.fr/5xzn0b

 

[lexluthermiester]

Oh, you may or may not belive me, but this screenshot was done on R9 390X (card isn't visible and can't be installed because of "Code 15*" error with PCI-e bridge [*System Firmware doesn't have info needed to make it work] :/).

Use the Windows Vista/7 drivers from the AsRock webite. They will work in Windows 10.

 

[agent_x007]

Drivers aren't the problem and OS either.
There are no "VGA adapters" in Device Manager.
As in : I do not have a device that driver can be installed for.
Interesting, isn't it ?

Installing chipset driver won't work either.
Because problem isn't driver, but MB's BIOS that doesn't understand how to operate this GPU (390X).
Funny thing is that it works with GTX 780 Ti without any problems

 

[lexluthermiester]

MB's BIOS that doesn't understand how to operate this GPU (390X).

Wow, that's weird!

Funny thing is that it works with GTX 780 Ti without any problems

Interesting. Reminds me of a similar problem that happened with old QDI boards back in the Pentium 2 & 3 days.

 

[agent_x007]

Wow, that's weird!

I know !
But that's what it's saying :

I thought this was because R9 390X dropped Legacy BIOS support in vBIOS, but (from what I understand) that would mean it woudn't work at all (ie. no POST) and a reflash would be needed.
However it clearly does work in basic VGA mode...
Like you said - WEIRD.

PS. No Video Adapters :

 

[delshay]

Drivers aren't the problem and OS either.
There are no "VGA adapters" in Device Manager.
As in : I do not have a device that driver can be installed for.
Interesting, isn't it ?

Installing chipset driver won't work either.
Because problem isn't driver, but MB's BIOS that doesn't understand how to operate this GPU (390X).
Funny thing is that it works with GTX 780 Ti without any problems

I have still not used the full speed of FX-60 with R9 Nano "1080p Extreme" https://www.techpowerup.com/forums/threads/the-fury-x-owners-club.214557/page-4.

And here's my ASRock 939Dual-VSTA in action : Windows 10 x64 on Socket 939 motherboard

Even LAN works (after Windows Update searched online for driver, because Smartphone + USB Internet Sharing = Internet)
Oh, you may or may not belive me, but this screenshot was done on R9 390X (card isn't visible and can't be installed because of "Code 35*" error with PCI-e bridge [*System Firmware doesn't have info needed to make it work] :/).
Here's valid : https://valid.x86.fr/5xzn0b

That's not 939, you can clearly see DDR2 memory & FX62.

 

[agent_x007]

That's not 939, you can clearly see DDR2 memory & FX62.

Just because you see "FX-62" and "DDR2", doesn't mean my board isn't Socket 939.
Let's just say 939Dual-VSTA (as seen in Mainboard CPU-z tab), is a little special in this regard

 

[delshay]

Just because you see "FX-62" and "DDR2", doesn't mean my board isn't Socket 939.
Let's just say 939Dual-VSTA (as seen in Mainboard CPU-z tab), is a little special in this regard

Well that explains it. I have seen these motherboards but never look into it. Why are running the processor at that voltage, it should work at 1.35v or lower if you undervolt. Mine go's as low as 1.75 - 1.2v at standard speed 2.6GHz.

 

[Trekkie4]

Wow, that's weird!


Interesting. Reminds me of a similar problem that happened with old QDI boards back in the Pentium 2 & 3 days.

Never heard about this issue before, would you mind sharing more info (or resources?). I got two QDI Advance mobos here, one with Pentium II (Intel 440BX) & the other one with Pentium III (VIA Apollo) & so far didn't have any issues. At least none that I've noticed *shrug*

Talking about Slot-1, let's discuss numbers & benchmarks...


Finally got around to finish that PII 266 build. It was meant to be exact copy of my Celeron 333 system, but due to uhh ... unforseen circumstances I ended up doing entirely fresh installation on Celeron 333. It also got a new cable management, CD-ROM drive, network card, additional 64MB of RAM, etc. The list goes on & on but either way - once I was done with Celeron 333, I dumped the entire hard drive on PII 266 system, since they're basically identical. Same mobo, same sound card, same amount of RAM. They only got different GPUs & CPUs, along with dial-up modems & CD-ROM (vs DVD on PII) I like to think of these as Data & Lore ... for those of you who don't get it, it's a Star Trek reference


On one end we have Celeron 333, paired with S3 Savage4 Pro (AGP) card, and the other one is of course Pentium II 266, paired with nVidia TNT2/M64 & single Diamond Monster 3DII, 12MB. What I find interesting is that M64 outperformed Voodoo2 by a small margin. I'm sure the CPU plays important role in all this, being the bottleneck for BOTH nVidia & 3dfx. But still ... fascinating nevertheless! Here are the actual numbers, for both of the cards:


Now, you may be wondering - OK, but how does PII 266 & M64 compare against (in)famous Celeron, running at 333MHz & S3 Savage4 Pro? The answer is almost identical, although slightly better. If it wasn't for crappy S3 card, this Celeron would probably do much, MUCH better!


By the way, we recently discussed nVidia's drivers & related issues. Here are the ones used for this project, ForceWare 29.42. Personally, I find 43.45 to be somewhat better (more reliable) but with that being said, I didn't have any reasons to include 43.45 since TNT2 series was already supported under 29.42. And you know what they say - if it ain't broke, don't fix it!

 

[Robert B]

Today I managed to open the cartridge of the SLOT A K7-800MHz Thunderbird core CPU . Even after I've opened many slot A and slot 1 CPUs this task is still full of stress and cold shivers on my spine

The die of the K7-800MHz sports a 850MHz core - not bad - A850CPRFFA - I WISH I HAD A GOLDFINGER DEVICE....



The capacitors have arrived

1. 22uF/16V, SMD, Panasonic EEE1CA220SR
2. 3300uF/6.3V, Radial, LOW ESR, 10x20mm, Nichicon UHM0J332MPD1TD
3. 10uF/16V, SMD, Low Esr, EEEFP1C100AR PANASONIC
4. 22uF/35V, SMD, Low Esr, EEEFP1V220AR PANASONIC
5. IC extractor D-PZ12



I have so many project going at once it is kind of ridiculous ...



More later.

 

[lexluthermiester]

Never heard about this issue before, would you mind sharing more info (or resources?).

Finding resources is dubious at best after 20years, but the general problem was if you matched a certain couple of cards with certain QDI mobo's the system wouldn't post or wouldn't show the card in the device manager in Windows. That situation was one of the earliest examples of a system needing a bios update, after which the systems ran fine. IIRC it had something to do with AGP compatibility.

I got two QDI Advance mobos here, one with Pentium II (Intel 440BX) & the other one with Pentium III (VIA Apollo) & so far didn't have any issues. At least none that I've noticed *shrug*

I don't think you're likely to have any issues. Most of those mobo's were either bios updated or sent back to QDI. You shouldn't have to worry.

The die of the K7-800MHz sports a 850MHz core - not bad - A850CPRFFA - I WISH I HAD A GOLDFINGER DEVICE....

You can FSB OC those CPU's. IIRC they had a ton of OC headroom. An old friend got his 750, which I think was a Pluto core, to 1ghz with a slight voltage bump. Bumped the FSB to 133 and leave the multi as is. Helped him custom mount a large heatsink and case fan to it for cooling. EDIT; Yup, the Argon cores stopped at 700, so his 750 was a Pluto. Same with your 800, and you can bump it to 133FSB and drop the multi down to 7.5 to get 1ghz. Even though it's an old CPU there shouldn't be any ill effects from such an OC. Those things were/are hardy and durable.

 

[dorsetknob]

Bought a QDI P3 Capable Board from New ( With a S370 P3 800Mhz CPU ) ran AGP Graphics Cards in it never had any problems never needed to update bios.
Maybe it was the AMD CPU Variants that were problematical ?

 

[Mr.Scott]

You can FSB OC those CPU's. IIRC they had a ton of OC headroom. An old friend got his 750, which I think was a Pluto core, to 1ghz with a slight voltage bump. Bumped the FSB to 133 and leave the multi as is. Helped him custom mount a large heatsink and case fan to it for cooling. EDIT; Yup, the Argon cores stopped at 700, so his 750 was a Pluto. Same with your 800, and you can bump it to 133FSB and drop the multi down to 7.5 to get 1ghz. Even though it's an old CPU there shouldn't be any ill effects from such an OC. Those things were/are hardy and durable.

That is not exactly true. There is no PCI lock on those boards and pending the board, most did not have adjustable straps for PCI/AGP bus speed.
Not to mention, the cache divisor suffers the same fate.
They're not as easy to overclock as you think.
GFD makes it easy. All you have to do is find one. It only took me 10 years.

An old friend got his 750, which I think was a Pluto core, to 1ghz with a slight voltage bump. Bumped the FSB to 133 and leave the multi as is. Helped him custom mount a large heatsink and case fan to it for cooling.

By the way, here is the records for 750 Pluto.
Make sure you check the WR frequency out.
http://hwbot.org/hardware/processor/athlon_750_pluto_slot_a/

 

[DR4G00N]

That is not exactly true. There is no PCI lock on those boards and pending the board, most did not have adjustable straps for PCI/AGP bus speed.
Not to mention, the cache divisor suffers the same fate.
They're not as easy to overclock as you think.
GFD makes it easy. All you have to do is find one. It only took me 10 years.

Isn't it possible to just make a GFD? I believe that all you'd need to know are the VID & FID tables and the pin-out to put one together. Finding a connector to fit on it may be the hard part if it has some odd pin pitch design.

 

[phill]

I've so much love for this thread... Just out of interest, how old is the nostalgic hardware up to??

 

[Mr.Scott]

Isn't it possible to just make a GFD? I believe that all you'd need to know are the VID & FID tables and the pin-out to put one together. Finding a connector to fit on it may be the hard part if it has some odd pin pitch design.

This might possibly interest you.
http://www.cpu-world.com/forum/viewtopic.php?t=28887

 

[agent_x007]

I've so much love for this thread... Just out of interest, how old is the nostalgic hardware up to??

I'd say... up to LGA 1366/AM3.


Pretty good for a Single Core

 

[stinger608]

This might possibly interest you.
http://www.cpu-world.com/forum/viewtopic.php?t=28887

Holy crap Scott!!!!! That's an awesome link man.

Any progress on the unit in them forums? Not a post since December 2017 so I'm assuming the builder has been busy with other projects.