lazibayer wrote:

Sometimes I wonder how Intel survived around the Millennium. Hyping on RAMBUS was a disaster, 820 chipset with RDRAM->SDRAM converter was a disaster, 845 chipset with single channel SDRAM was a disaster, and the biggest among all, the early Pentium 4.

Don't forget the Itanic 😁

Intel has more legs to stand on.
They sell lots of other IC's and CPU's. Also the 80386 was made until 2007 😉

Which detonators are you guys using ?

sunaiac wrote:

Which detonators are you guys using ?

I have been using 92.91 in Windows XP with the Ti 4200. 92.91 is probably unnecessarily new and not the fastest driver for a GF4 Ti 4x00 but it's a driver known for working with just about anything. 😀

I have not seen any glitches or noticeably bad performance.

In Windows 98 I think 45.23 is the driver to use! 😀

Tonight and probably tomorrow I need to test the Asus A8R32-MVP DELUXE socket 939 motherboard with CPU and memory I just bought and also a GTX285 I also bought to see that everything work so I can leave feedback on those items.

I will be back to the K7 tinkering / benching this weekend.

matze79 wrote:

Also the 80386 was made until 2007 😉

do You happen to have any details on that? what chip was that exactly and used how? how would i know when im looking at such a modern 386 chip?

Spend one hour installing XP sp3. Check.
Have the hard drive die just after the last reboot. Check.

Sometimes... Rhaaaa !
I guess I'll just order a large compact flash.

sunaiac wrote:

Spend one hour installing XP sp3. Check. Have the hard drive die just after the last reboot. Check. […]
Show full quote

Spend one hour installing XP sp3. Check.
Have the hard drive die just after the last reboot. Check.

Sometimes... Rhaaaa !
I guess I'll just order a large compact flash.

I recommend a SD-Card to IDE adapter. They cost only a few $ more, but you can make that back because SD cards are in much easier supply. Currently installing Win98SE on a 2GB SD to test it out.

Also... I have an issue with my A7N8X not saving BIOS settings after a power shutdown (G3 state). I thought it was the CMOS battery, but I put in a brand new one, and it still does not save the settings. Also checked the 'Clear RTC' jumper, and it looks to be default. Any ideas?

If you guys are interested, I'd like to see some GUI performance tests. I suggest the Tom2D benchmark. I did some runs a few years ago because I'd noticed that VIA chipsets always felt sluggish compared to NV NForce. Web browsing and just minimizing/maximizing windows come to mind. The results showed a considerable GUI performance deficit on the VIA platforms, across a few different GPUs. Even the K8T800 platform was slower than nForce2.

swaaye wrote:

If you guys are interested, I'd like to see some GUI performance tests. I suggest the Tom2D benchmark. I did some runs a few years ago because I'd noticed that VIA chipsets always felt sluggish compared to NV NForce. Web browsing and just minimizing/maximizing windows come to mind. The results showed a considerable GUI performance deficit on the VIA platforms, across a few different GPUs. Even the K8T800 platform was slower than nForce2.

The download link at tomshardware.de seems dead. I think the program is small enough to attach as an attachment.

Skyscraper wrote:

The download link at tomshardware.de seems dead. I think the program is small enough to attach as an attachment.

Here you go

Toms-2D-Bench
http://www.mediafire.com/file/op7h65rvubj3atn … ms-2D-Bench.zip

swaaye wrote:

Here you go […]
Show full quote

Skyscraper wrote:

The download link at tomshardware.de seems dead. I think the program is small enough to attach as an attachment.

Here you go

Toms-2D-Bench
http://www.mediafire.com/file/op7h65rvubj3atn … ms-2D-Bench.zip

Thanks

Last weekend I noticed a difference between nForce2 and VIA KT880! 😀

With nForce2 you always get the best performance by running the memory in sync with the FSB but with the KT880 memory speed is king even if using a divider. I learned this by forgetting to change the memory divider. The difference is small but I re ran 3dmark 2001 a few times and running the memory using the 3:4 divider comes out on top more often than not.

Athlon XP 3200+ (Barton 14*166), DC memory at 166MHz DDR CL 2-2-2-6 1T, Ti 4200. 3dmark 2001.

Athlon XP 3200+ (Barton 14*166), DC memory at 222MHz DDR CL 2-2-2-6 1T, Ti 4200. 3dmark 2001.

This weekend I plan to test my Abit AN7 nForce2 motherboard a bit as we have not really finalized the benchmarking setup and what benchmarks to use yet (it's no hurry).

Before seeing what the Abit AN7 could do I had to fix the missing chipset cooler fan. It's was probably good that the fan was missing as the stock north bridge cooler didn't make very good contact with the chipset at all. I replaced the stock cooler with some heatsink I found and added a small fan.

I have messed a bit with the motherboard already and I have to say that I like it alot! When reading old forum posts one gets the impression that the Abit AN7 won't do FSB speeds over 215 MHz without disabling both CPC and "CPU interface" and not FSB speeds over 225 MHz without voodoo and blood sacrifices. My experience is luckily very different.

Got a cheap asus a7n8x deluxe with a 3200 barton

Since I want to oc it, so water cooling is preferable but hard to find. But I noticed the 4 holes which a standard socket 462 cooler doesn't use. Anyone know what kind of socket (water)cooling can I put on this mobo?

I'm thiking about some graphs with CPU scallng and GPU scalling with 3DMark 2001 SE using all my K7 CPUs, multiple graphic cards for the GPU scalling. Also planning to add some Pentium 3/4/Celeron (K7 competitors) and old some Socket 754 AMD Athlon 64/Sempron (CPUs supossed to replace K7) CPUs for comparasion pruposes in the CPU scalling graphs. It might take some time to start working on it and make it happen though, i'll try to get a faster GPU than my Radeon x1650 Pro AGP. OS will be Windows XP SP3 for all the system/CPUs, hardware is still undecidded

I have had a buisy week but I have at least had the time to so some more testing with the Abit AN7... or more like trying to do some testing with the Abit AN7.

My now somewhat complicated relationship with the Abit AN7 started out great last weekend. I could run my Athlon XP 2800+ at 11x200 MHz with 1.5V and 12x200 with 1.7V and I could also run 10x240 with 1.7V and default chipset voltage. 240 MHz FSB seemed totally stable with all performance tweaks active including CPC but with relaxed 3-3-3-8 timings as this was with a single infineon 1GB stick. I wanted to see what the motherboard could do when running dual channel and tighter timings.

I used a couple of very capable 512MB Winbond BH5 sticks and running dual channel at 200 MHz with 2-2-2-5 timings and CPC was not an issue but anything faster than that wouldn't stay 100% stable regardless of settings. I tried alot of other memory sets with differnt types of memory chips but the stability issues only seemed to get worse. The instability first only showed at high overclocked settings or with overclocked memory runnings in Dual Channel mode but later it became an issue even at stock settings and in the end even at underclocked settings regardless of memory configuration and CPU used.

I thought that something had become corrupted so I reflashed the BIOS and reinstalled everything, nothing worked and the motherboard was now not even close to stable even at fail safe defaults. At this point I thought F**K IT and throwed the motherboard into the dishwasher without removing anything but the battery, I even left the BIOS chip in its socket. My idea was that a full program should sort out any gunk in the CPU socket, AGP slot or memory slots as the motherboard was really filthy when I got it.

After the dishwasher treatment the motherboard looked like new! I used the oven to dry the motherboard quickly, not cap friendly but I wanted it dry ASAP. After sweating for 30 min at 100C, 30 min at 75C and then 30 min at 50C I thought the motherboard was dry enough and tested it, the instability was still there. I thought about giving up at this point but something made me try to increase the CPU voltage. It turned out that the motherboard works fine as long as I use 1.85V core voltage or more...

I have tried different PSUs and I have measured every voltage including the core voltage, I can not find any reasons for the sudden instability at lower voltages. At least the issue does not seem to get any worse at this point and CPU and memory overclocking still work fine but of course only as long as I use at least 1.85V core voltage. The same 1.85V is needed to get stability at 11x166 (1833) and 11x228 (2500 MHz).

I have now read about other people with similar issues with the Abit AN7 and the solution (from what I can gather) could be to reflash some really old BIOS before again flashing one of the newer and/or tweaked BIOS versions. I have yet to try this as overclocking seems to work great at the moment and that is what I wanted the Abit AN7 for anyhow.

Some performance numbers at 2500 MHz. I'm still using the small Akasa cooler so this is as high the cooling lets me go.

Abit AN7, Athlon XP-M 2800+ @ 11x228 (2500MHz), 1x1GB DDR @228 2.5-3-3-11 CPC. Super PI 1M: 41s

Abit AN7, Athlon XP-M 2800+ @ 11x228 (2500MHz), 1x1GB DDR @228 2.5-3-3-11 CPC. 7-Zip 32M

Abit AN7, Athlon XP-M 2800+ @ 11x228 (2500MHz), 1x1GB DDR @228 2.5-3-3-11 CPC. PCMark 2002

Abit AN7, Athlon XP-M 2800+ @ 11x228 (2500MHz), 1x1GB DDR @228 2.5-3-3-11 CPC, Ti 4200 128MB. 3dmark 2000

Abit AN7, Athlon XP-M 2800+ @ 11x228 (2500MHz), 1x1GB DDR @228 2.5-3-3-11 CPC, Ti 4200 128MB. 3dmark 2001 (This is how a bottleneck looks like!)

Lets see if running high FSB using single channel memory and relaxed timings is faster than running 200 MHz FSB with dual channel memory and tight timings.

First out is the setup running at 200 MHz FSB with dual channel CL2-2-2-11 timings.

Abit AN7, Athlon XP-M at 12x200 2x512 MB Dual Channel CL2-2-2-11 T1. Super PI 1M: 42s

Abit AN7, Athlon XP-M at 12x200 2x512 MB Dual Channel CL2-2-2-11 T1. 7-Zip 32M

Abit AN7, Athlon XP-M at 12x200 2x512 MB Dual Channel CL2-2-2-11 T1.PCMark 2002

Abit AN7, Athlon XP-M at 12x200 2x512 MB Dual Channel CL2-2-2-11 T1, GF4 Ti 4200. 3dmark 2000

Abit AN7, Athlon XP-M at 12x200 2x512 MB Dual Channel CL2-2-2-11 T1, GF4 Ti 4200. 3dmark2001

Lets see if we can beat the score above by by clocking the CPU to 10x240 using single 1GB memory module running at more relaxed timings.

Abit AN7, Athlon XP-M at 10x240 1x1024 MB Single Channel CL3-3-3-8 T1. Super PI 1M: 42s

Abit AN7, Athlon XP-M at 10x240 1x1024 MB Single Channel CL3-3-3-8 T1. 7-Zip 32M

Abit AN7, Athlon XP-M at 10x240 1x1024 MB Single Channel CL3-3-3-8 T1. PCMark 2002.

Abit AN7, Athlon XP-M at 10x240 1x1024 MB Single Channel CL3-3-3-8 T1, GF4 Ti 4200. 3dmark 2000.

Abit AN7, Athlon XP-M at 10x240 1x1024 MB Single Channel CL3-3-3-8 T1, GF4 Ti 4200. 3dmark 2001.

The difference isn't huge but yes, running at a high FSB using single channel memory is faster than running at a lower FSB using dual channel memory.

How is the system building for benching progressing? I have to admit that my preparations has stalled a bit because I got sidetracked by some fun Slot 1 stuff I had to tinker with and by work and life in general but I could be up and running within a day or two.

Finished Benching with my new Barton.

Asus A7N8X Deluxe
AMD Athlon "Barton" XP 3200+ & Thermaltake Volcano 12 cooler (modded with a 120 mm instead of the stock one)
2 Gb Dual channel 3-3-3-8
Asus A9800 XT
All at 1024*768 and standard testing:
3D mark 2001: 16175
PC mark 2002: CPU 6826, Memory 6605, HDD 1192
3D Mark 2003: 7142
3D mark 2005: 3407
7-Zip rating: 1744

Temperatures while stressed are not to shabby although I am not happy with the northbridge temperature measurements, so before I start the OC on this build I need to get that northbridge in check (might aswell do the southbridage aswell at the same time with some thermal pads and a cooler.
Stressed temperatures after 15 minutes: with prime95 to stress the CPU

Edit: changed the stock passive cooling to a GPU heatsink (30x30x10) and a cooler, this decreased idle temp with 10 degrees. Stresses temp is still around 60 degrees max ( so a 5 degrees decrease). Either Increase air flow or get a adapter for a bigger fan on the northbridge...

the voltages were very satisfactory aswell (Hiper modular 580W), when stressed voltages remain around these numbers with ocassional spikes still within the acceptable borders for a PSU.

Are all Barton Athlon XP-M's clock unlocked? I remember that the XP-M 2500+ was quite popular among overclockers.

And lastly, can anyone recommend a decent Socket-A heatsink that doesn't require pressing down with a flathead screwdriver? I a Socket-A heatsink I pulled from a Compaq system that uses a lever to put force on the heatsink, but I need something a little better quality.

dexvx wrote:

Are all Barton Athlon XP-M's clock unlocked? I remember that the XP-M 2500+ was quite popular among overclockers.

And lastly, can anyone recommend a decent Socket-A heatsink that doesn't require pressing down with a flathead screwdriver? I a Socket-A heatsink I pulled from a Compaq system that uses a lever to put force on the heatsink, but I need something a little better quality.

During my OC run yesterday I squeezed 24xx MHZ out of my 3200+ Barton, took abit of effort and stability is meh. Temps didn't exceed 50 degrees on the cpu and I am still working on better cooling inside the case (north bridge was overheating due poor airflow)
After those tests I believe the 3200 Barton is not very suited for OC higher then 200-250 Mhz. Specially combined with the nforce chipset, the kt880 might run cooler and more stable along with a Barton.

Edit:
Due lack of time I couldnt run a comprehensive benchmark on all the OC. With the gpu OC boost I squeezed out 3xx points on 3dmark2003. Will try to get a full list this weekend as the additional score got me excited and I want to run quake 4 on higher settings stable 🤣

I currently use the volcano 12 from thermaltake which also uses the lever to press down on the heatsink, I searched quite a while for a decent one. If there were watercooling adapters for this socket I would rather use watercooling due weight limit of the die.

edit2:
finally got around for the benchmark:

also finally solved my Nforce chipset temperature problems after adding 2 (maybe 1 was enough but had room for 2 fans) exhaust fans to the side of case. it went from roughly 60 ish degrees on average to 30 degrees (replaced cool paste and added a 30 mm heatsink & fan

Well I think I messed up my A7V333. Today, my AXP 2600+/266 Thoroughbred (fastest 266 FSB) arrived. I removed my AXP 2000+ Palomino, put the 2600+ and started to mount the heatsink/fan (the ones where you have to press down with a flathead screwdriver). Screwdriver slipped, hit the motherboard (not even that hard, IMO). Now my setup won't even post 😵 with anything. I know both CPU's are good because they POST on an A7N8X.

Sad thing is, I don't even see any physical damage on the A7V333 (I mean obviously there is).

Can anyone recommend any Socket-A heatsinks that don't require a flathead screwdrive? TT Volcano's seem kinda rare.

dexvx wrote:

Well I think I messed up my A7V333. Today, my AXP 2600+/266 Thoroughbred (fastest 266 FSB) arrived. I removed my AXP 2000+ Palomino, put the 2600+ and started to mount the heatsink/fan (the ones where you have to press down with a flathead screwdriver). Screwdriver slipped, hit the motherboard (not even that hard, IMO). Now my setup won't even post 😵 with anything. I know both CPU's are good because they POST on an A7N8X.

Sad thing is, I don't even see any physical damage on the A7V333 (I mean obviously there is).

Can anyone recommend any Socket-A heatsinks that don't require a flathead screwdrive? TT Volcano's seem kinda rare.

any decent heatsink for performance is rare these days, although Amibay currently has a arctic cooler up for sale which is a good one. Add a 80 --> 120 mm adapter to it and you have a great cooler capable of heavy OC at very low noise levels.