• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

14/400 in 48 hours!!!!

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
This machine has been on 24/7 for the last 2 years not folding. It may as well be doing something productive!

Anyways, once I get settled into our new house I'm going to put together the rig in my sig.
 
That's going to be a pretty nice setup once you get it together. Although...a 550mhz CPU wouldn't be more than like $20.
 
I think the M570 can only do up to 75Mhz bus speed, so 400Mhz would be the max AMD CPU it could use. K6-2-333s are notoriously poor overclockers too :( I've only ever managed a barely stable 375 out of one with a hefty sink on it. K6-3s are of course a bit faster due to on die L2, and the plus models might just run gromacs because they have 3DNow pro, but unless you get a BARGOON on the chip it's not worth it, considering you can probably get a duron and low end duron board for $40 or less maybe.

Anyhoo, good luck with the folding apenland, keep at it, I run a few K6-2s as surfing boxes with folding on them and know your pain. It's quite a rush to see that 70 point boost every couple of weeks though. That's 35ppw, when I first started they'd fold maybe 3 or 4 6 pointers a week. I made my first 500 points on K6-2's alone.

Road Warrior
 
Apenland,

I cheer on your pow-ah!!! You K6 harkens back to the days of yore, when DC projects were usually crunched by left overs and frankenstien boxes.

If it can post, it should fold!

Wedo
 
This PCchips board is supposed to be a 100mhz board, but I have never been able to get it over 66 with any stability. My wife won't let me corrupt the data in the HD anymore, so no OCing with this baby. When I turn this box off to move, it may never start again. 6 months ago I cut the power to clean it up and when I started it back up it sounded like a train wreck and took 20 minutes to stop making the noises.

I'm reading these forums to see if it can fold out in the garage without being connected to the net directly. If it can, it will fold!

PS: 17/400
 
yeah fold on bro, as long as its for team 32 =P stil nice rig tha you are building
 
75 x 6.... 450Mhz.... all k6-2's can use 6x multi, just set the motherboard to 2x, it will think its 6x :D
 
apenland01 said:


I'm reading these forums to see if it can fold out in the garage without being connected to the net directly. If it can, it will fold!

You need to be connected to the net to send and recieve work units, couple nic's and a crossover cable would do it and be like less than $25

IF you can't go above 66FSB, hit it with a 6x multi, can u say 400Mhz ;)
 
What you can do, is if the garage is apart from the house, is set up a wifi card in it, and just have it connect to the net by that. That way you won't need to worry about running any wires. :D
 
JDXNC said:
75 x 6.... 450Mhz.... all k6-2's can use 6x multi, just set the motherboard to 2x, it will think its 6x :D

Not all. There are two versions of the K6-2, the original "Chomper" core, which was released in speed grades only up to 350, can only do a 5.5x max multi. The second revision, known as the CXT "Chomper eXtended Technology" has the 6x multi remapped from the 2x. It also has a couple of speed enhancements that some older motherboards don't enable, write combining being one, but you can run a util like "setK6" to turn them on. The greater part of K6-2-333s around are original chomper cores and are not 6x capable, it was originally thought that AMD was only releasing CXT in 400Mhz and above K6-2s, but that may have only been true for a few months to a year, they did release some lower speed grades, 380, 350 CXTs I have seen, I also have a 333Mhz CXT, this has a production date in 1999 stamped on it. If it's a 1998 chip chances are it's an original chomper and won't support 6x.

AFAIK they never ever released CXTs lower than 333 so 266 and 300 Mhz parts will definitely not have the 6x multi.

Some sites erroneously refer to the CXT K6-2s as sharptooth, the sharptooth is the K6-3, not the K6-2 CXT.

A tip for K6-2 stability at higher bus speeds on PC Chips boards is to set the ram timings lax, crazy glue some ram sinks to the onboard cache, set the multi to a .5 multiplier, set the bus speed to 83 with the 2/5 PCI divider that puts the PCI bus at 33Mhz. For some reason this async mode works best on a .5 divider, probably because the bus timings allow a wait state when set to .5 or something.

http://users.erols.com/chare/586.htm lists the early chompers and CXT seperately, the dates stated are announcment dates, I'm pretty sure that nothing other than a 400 or better was available as a CXT for a few months after those dates, and the 300 CXT may have been a paper release with demand for that speed collapsing before the old stock of chompers was sold out, besides them needing the CXTs more in the higher grades.

heh, sorry, more than you every wanted to know about K6-2s :D

Road Warrior
 
Erhm ... I hate to be a naysayer ... but ...

If my math is correct, that box will take 57 days to complete that WU. I just looked at the project summary page, and the longest deadline I can see for a 400 frame Tinker WU is 50 days.

IIRC, the beanchmark is set with a PII 400 MHz CPU ... yes?

And ... isn't the K6 a Pentium Classic equivalent? I can't remember exactly ... but even if it's a PII equivalent, it seems to be a bit slow for the task at hand.

Maybe you should task it for GAH.

Is Fry's Electronics in your area? If not, take a look at Outpost.com . You can pick up some great deals on MoBo/CPU combos.

... just my .02 - let the flaming begin! :D
 
Loud said:
And ... isn't the K6 a Pentium Classic equivalent? I can't remember exactly ... but even if it's a PII equivalent, it seems to be a bit slow for the task at hand.

Depends how you define equivalent. The original K6 core has a dual integer pipeline and branch target prediction like the PPro core, which the PIIs use, and it has MMX like the PII, whereas the PII has L2 cache on the processor module though at half speed, the K6 had it on the mobo still at 1/3 to 1/5 speed depending on the multiplier, the cache is synch with the FSB. The K6-2 takes things a step further with 3D now, which are SIMD streaming extensions almost the same as SSE, but still had cache on the mobo. I guess you could call the K6-2 a P2.5 equivalent though :D The K6-3 brought the cache on die at the full core speed and thus outclassed the PII and competed on level ground with early PIIIs. However with the newer super 7 motherboard capable of running 124 or even 133 FSB, K6-2s generally can hold thier own against a PII, performance that was not possible "back in the day" with EDO RAM or SDRAM limited to 66Mhz.

The benchmark machine they use for F@H (Which I heard was a 500Mhz celly btw) is taken at 8 hours a day run time. So theoretically anything that performs only a third as good as a celly 500 but folding 24/7 has a chance.

Road Warrior
 
RoadWarrior said:


If it's a 1998 chip chances are it's an original chomper and won't support 6x.



Road,

I got this system in December 1998, so it's a 98 chip. I have 66,75,83 and 100 for choices and I can't seem to get anything above 66. 75 works well for a few days, then data corruption. Once I get all the data transferred to the new system HD, I am going to follow some of the tricks you list here since this box will be for the boy to play with and corruption won't be so "wife problematic".

This box folds for ~21 hours a day, so it may make it as I have until April 6 for this WU.

Currently 24/400!!
 
Back