If you intend to use it on a 90º position, I hope you don't break your socket... That weights a lot.

How much does half a kilo of copper go for these days?

I used one of those ones that looked like an air turbine. Had the pc standing as a tower. When I removed the heatsink, it was dented in where the corners of the cpu pressed. After that I went back to normal heatsinks, I didn't want to risk damaging the cpu die.

Some of the heavier heatsinks bolt through the board; of course they need to have holes. Still must put a lot of pressure on the pcb though.

Behold - The Ripper! 🤣

Be careful with that one. It looks as if it could do nasty things to the CPU and socket.

I'd be worried about it ripping off the board if it's not bolted through.

Also the fan mount looks like it blocks up to 20% of the fins!

Big heatsinks are common for Intel Core i5/i7/i9 CPUs these days, but not sure about Athlon XP. I think the one on my Athlon 64 X2 is relatively small.

It should be ok with it using all three socket prongs.... But that baby must weigh a bunch. Usually they make the fins of tower sinks out of Al.

I have just installed The Ripper! I wonder how many "lives" it has claimed? Actually, I am pleased with it. I was very careful when installing it. Took about 3-4 mins to coax it in to position. It's on an Epox EP-8KTA3+Pro board. I increased the FSB to 152, giving the 3000+ rated CPU a real over clocked speed of 2510 MHz. Quite speedy.

I ran SuperPi.exe at the Windows 98 desktop. I ran the 1M test, twice in a row. Then I queried Sandra 2002 Pro for the CPU temp. It was 33.5C. Not bad! I ran coolon.exe, and then reran SuperPi's 1M test, twice. The temp was 33.4C. That's really interesting - Coolon didn't appear to help lower the temp. Well wpcredit is working OK, because I used CPUMSR.exe to set the CPU's multi to 16.5x.

BTW, for the SuperPi.exe 1M test, I get a result of 000h 01m 06.594s. Interesting tidbit of info: although I'm using 2 PSUs, the one connected to the mobo is a bit crap. It's only 350W, and crucially the 5v rail is only 30A, and its reading in Sandra is 4.76 for the 5V rail. But the system seemed stable. Also, the heatsink's 80mm fan runs at 1800 RPM, and is very quiet.

Wow that thing is enormous. I would definitely be concerned about the socket. If you install the motherboard in a case, the cooler is going to put a lot of pressure on the topmost two prongs with that socket orientation.

I didn't enter the realm of tower coolers until Athlon 64 and they had redesigned the mounting mechanism into something much more resilient by then.

At the moment, I don't think I'm ever going to install the mobos I test in cases. I've probably got years more testing to do before I ever reach the "end product".

I've got some more copper coolers to try. I'll test them ASAP, and then report back with the temps. The smallest one on the left is slimline in shape, and fits on (super) socket 7 mobos!

I used to have one of those Zalman peacock-style heatsinks (the centre one)on an Athlon XP 1800+ Palomino. It was pretty quiet and kept things reasonably cool although there were definitely better performing heatsinks out there.

What a beast!

My tip is to remove / break off that plastic casing around the cooler and mount the fan directly onto the copper cooler. Either with wood screws, cable binder or some other method.

Yes that's a good idea to remove the plastic fan mount, because it's blocking some of the fins.

I tested the 3 other heatsinks. I also retested "The Ripper". This time, I ran Prime95 in the background, then ran the SuperPi 1M test twice. I get these temp readings:

30.7 Homight "The Ripper"
34.4 Coolermaster "Heat pipe"
35.3 CoolJag "Slim line"
37.1 Zalman "Peacock"

For the Zalman and Coolermaster tests, I used a 1875 RPM fan, because the other heatsinks had their own fans. Some other observations - with Prime95 running in the background, this made absolutely no difference whatsoever to SuperPi's score. Strange that. Also, SuperPi's score was unaffected by any memory timing setting adjustments made inside the BIOS.

Edit: I'm really intrigued why Coolon.exe made no difference to the temp. I am using a different PSU to the one that I normally use. Maybe Coolon.exe is dependant on what type of PSU you are using? It's a new 350W PSU, with an unusual -5V rail. It's unusual because it is a brand new manufactured PSU, made by Nexus, and new PSUs do not usually have anything on the -5V rail.

Mainboards have no knowledge of the PSU type.

h-a-l-9000 wrote:

Mainboards have no knowledge of the PSU type.

OK, thanks for the info. I wonder if Coolon.exe has no effect on the CPU temperature, if I use a particular CPU? The CPU which appears to make Coolon.exe not function is a 3000+ rated mobile barton. In the past, I have used a 2400+ rated mobile barton, and I can see that Coolon works OK, because the temperature definitely drops. This problem is not too important. It's just interesting.

Maybe coolon works by lowering the multi/voltage of mobiles but not for desktop XP's? All I can think of from the top of my head

Tetrium wrote:

Maybe coolon works by lowering the multi/voltage of mobiles but not for desktop XP's? All I can think of from the top of my head

I'm afraid I don't know how Coolon.exe works. However, the two CPUs I have tested with Coolon are both mobiles; a 2400+ and a 3000+. The 2400+ worked with it, but the 3000+ did not. I unlocked the 2400+ by cutting a bridge on it, but I have not unlocked the 3000+ yet. Could that be the reason, I wonder? Do I need to unlock the 3000+ CPU? (Actually, I doubt it.) Maybe the 3000+ chip is simply "incompatible" with how Coolon works.

I'm happy with the performance of the overclocked 3000+ chip, in conjunction with most of the big copper coolers. 😀 Coolon apparently not functioning correctly is not a major problem.

Say Tetrium and Mau1wurf1977, what hardware have you bought/got recently? Anything exciting? 😀

Except for the 9 P1's, not much actually. Been busy with other stuff lately. I did buy some other chips which still need to arrive here (couple P2, hopefully unlocked, a 486 board untested, couple P1-200 overdrives).
I did buy a few odds and bits locally, like an external cdrom drive which I bought mainly for the housing and because it uses a standard cable to connect directly to the wall. Haven't tested any of that stuff though, been occupied with other things right now.

Edit: The extrenal cdrom drive is btw parallel I think.

Coolon changes the chipset register setting to enable the HLT instructions. It's the same thing you have done with wpcredit.