600E = 650E?

[-lynx-]

i mean for overclocking purposes? Just hike the FSB up and the same effect only the multiplier at 6.5 rather than 6, right? (please-please be that - I might lay my trembling hands on a 650E Slot1 but 600s are becoming extinct very rapidly)

------------------
-lynx-
13 Sqn RAF

[sparkzz]

For all purposes its the same.  You may have problems getting to 133 fsb with a 6.5 multiplier.  The coppermines seem to pilfer out around 850-950mhz without extreme cooling methods so its a luck of the draw issue getting to 900mhz.  I have a 600e running rock stable at 810mhz, but the heat from the chip is really high.  Get a really big heatsink, up the voltage a little, and watch your framerates soar!  

[Bombjack]

What he said  

[NUTTZ]

 

Originally posted by -lynx-:
i mean for overclocking purposes? Just hike the FSB up and the same effect only the multiplier at 6.5 rather than 6, right? (please-please be that - I might lay my trembling hands on a 650E Slot1 but 600s are becoming extinct very rapidly)

the 600E has 256 cache 100 bus
the 600EB is 512 cache and 133 bus
From my understanding the 600EB cannot be overclocked ( well it can, but i was told NOT to try it)
the extra cache on the chips will far out perform an over clocked chip with lesser cache, and be rock solid, and live to it's lifetime.

NUTTZ

[sparkzz]

Nuttz your confusing the Katmai specs with the coppermine designations.  There is to date NO coppermine chip with more than (or less than for that matter) 256k cache.  The E,EB designations are used to separate the new .18mic chips from the older Katmai .25mic chips.  The E and EB chips are identical in layout with the exception that the E uses 100mhz fsb and the EB uses 133mhz.
Both are overclockable, but your success with the EB will be much lower.  Example:
 
600EB (133fsb, 4.5 multiplier) with high quality PC133 (very expensive) and an excellent overclockers motherboard (expensive again) might reach 150 fsb.  That yields 675mhz.  The system will be flaky and prone to Blue screens, or the worse black screens (i.e. you couldnt even boot far enough to get a blue screen).  

600E (100fsb, 6.0 multiplier) with the same setup as above running 150 fsb would yield 900mhz.  More importantly though (since most of us cant afford the high dollar hardware) the 600e can be easily overclocked using a decent motherboard and economy PC133 (the cheap stuff) to 133mhz which yields 800mhz.  

It should be a clear cut decision which way to go.  (600E.. HINT: 600E) Your notion of cache size is correct. Size does matter.  Chips with same speed, but one running more cache will out benchmark the other.  But when one chip is running 800mhz and the other is only running 600mhz all the cache in the world isnt going to help.  (OK, it might, but can you imagine the size of the chip)  

[NUTTZ]

Given "my" choice i opted for Not overclocking, going for more caching, of course i got a good mom board that was ATA and a 7200 rpm HD, and 4x AGP for the GeForce.. What i, myself wanted was reliability AND speed. You often hear of overclocking( i too have gone that route ) but the Lifetime of the chip  AND added expence of Cooling the chip was factured into my final desision of the 600EB (512 cache on the chip),,,, let me tell you with the Geforce, Asus mom board, running at 133 bus AND 512 133ram, my frame rates are stable at 103 ( 800x600, 32 bit mode) sorry if i was confusing with the 2 different types of chips, comparing apples and oranges. But my performance in this box is stable , outstanding, and i know will last awhile ( without have to look down occasionally to see if my comp is smoking) Just another POV, even when price is factored in the 600eb series is pretty cheap.
NUTTZ


 

Originally posted by sparkzz:
Nuttz your confusing the Katmai specs with the coppermine designations.  There is to date NO coppermine chip with more than (or less than for that matter) 256k cache.  The E,EB designations are used to separate the new .18mic chips from the older Katmai .25mic chips.  The E and EB chips are identical in layout with the exception that the E uses 100mhz fsb and the EB uses 133mhz.
Both are overclockable, but your success with the EB will be much lower.  Example:
 
600EB (133fsb, 4.5 multiplier) with high quality PC133 (very expensive) and an excellent overclockers motherboard (expensive again) might reach 150 fsb.  That yields 675mhz.  The system will be flaky and prone to Blue screens, or the worse black screens (i.e. you couldnt even boot far enough to get a blue screen).  

600E (100fsb, 6.0 multiplier) with the same setup as above running 150 fsb would yield 900mhz.  More importantly though (since most of us cant afford the high dollar hardware) the 600e can be easily overclocked using a decent motherboard and economy PC133 (the cheap stuff) to 133mhz which yields 800mhz.  

It should be a clear cut decision which way to go.  (600E.. HINT: 600E) Your notion of cache size is correct. Size does matter.  Chips with same speed, but one running more cache will out benchmark the other.  But when one chip is running 800mhz and the other is only running 600mhz all the cache in the world isnt going to help.  (OK, it might, but can you imagine the size of the chip)  

[sparkzz]

Nuttz.. I dont know what to tell you.  You might want to check the Intel website for the specs on the 600EB.  It has ATC 256kb on die cache.  If you have 512kb you either have a rare engineering sample (not very likely) or a PIII 600A.  The "A" meaning .25mic and half speed off die cache.  But you would never get a 600A to 133mhz without some significant voltage hikes and extreme heat so I'll assume you do in fact have a 600EB.  You should post the chips stepping number.  Im really curious of its origin if it does in fact have 512kb.  Also, why so much memory?  Do you use the computer as a server?  The perfomance gains beyond 200-300mb of memory is dismal if not worse.  Usually that much memory only helps a server with significant workload.  The perfomance hit will be even worse on your Asus P3C motherboard since the i820 is not native to SDram and it uses a convertor hub to implement it.  The hub is well known to be a bottleneck that chokes the memory performance.  In fact the i820 perfoms worse than the Via133 chipsets which inturn perform worse than the Bx boards on the market.  Personally I chose the Abit BE6 II.  It has the onboard ATA66 and uses the faster BX chipset.  Oh well...  Like you said its a different POV. (apparently very different)  So I guess I'll keep my 600e running 810mhz rock-stable (without smoke either, does makes a nice foot warmer though) with my expensive cooling techniques.  (That AOC heatsink cost almost $15)  

[-lynx-]

Wow-wow! Calm down gentlemen!

The question actually was about using 650s (Es) instead of 600s for the very same overclocking routine. I dunno about the States but here in the UK 600s are plainly unavailable at the moment (well - at reasonable prices anyways) with no prospects of them appearing ever again with Intel launching new (=expensive) chips soon-ish.

I saw a few 650s here and there but all the "overclocking experiences" usually refer to 600E as a chip to wrestle with.

I'm not fussed about "getting it up to 133" either - 6.5 at 129 or even 124 still should give a massive hike in performance plus this would keep the AGP clock down a bit.

What I find confusing looking at my 450 currently clicking happily at 112FSB (+5C from what it used to have at 100, hence no extra heat-dispersing arrangements) is that the heatsink seems to be attached permanently to the whole thing - no obvious way to take it off without breaking something. When you guys refer to some better cooling arrangements I presume this means actually replacing what comes with the CPU as standard with something better, right?

Plus (running BX6 r2) I can play with quite a few CPU settings, including voltages etc. - "...up the voltage a little..." sounds cool but is there a reference to how little this "little" is actually supposed to be?

Thanks for all your help.


------------------
-lynx-
13 Sqn RAF

[Bombjack]

Sparkzz knows what he's talking about  

Lynx, the good thing about 133MHz is that at that point you can put the PCI divider to 1/4 and run your PCI cards and hard drives at exactly the spec speed. Above or below 133 and you're looking at something out-of-spec whichever way you cut it. 124's kind of mid-range - it might cause a problem.

The coppermines run pretty cool even with the stock heatsink/fan on. However I'm about to pick up some extra cooling for mine before I try and push it beyond 800 (got to offload my remaining PC100 memory too). I don't anticipate any difficulty at all doing that, but I haven't done it myself yet. Someone else can tell you the whys and wherefores.

Higher Vcore values can help CPU stability when overclocking at the expense of increased heat (which already be greater than normal at higher clock speeds). Standard Vcore for coppermines is 1.65v - if you've established that the rest of your system is happy at a certain clock speed (memory, AGP, PCI and so on all able to handle it) but the CPU is still occasionally letting you down, you might try incrementally bumping up the voltage, using the smallest steps your motherboard will allow. Last resort only though.

[sparkzz]

Removing a heatsink varies from chip to chip.  On your 450 (assuming PIII) it should be attached with 4 small pins or screws (seen both).  You have to remove the SECC casing to get access to these and that will void your warranty.  But if that doesnt bother you use a razor blade to wedge open the corners of the black casing and continue to work the razor blade around the casing's seam til removed.  The attachments to the heatsink are located on the backside of the SECC chip.  
As for replacement heatsinks the best ones come from Alpha and Global-win ($20-$40) with AOC making some knockoffs of both at marginal perfomance losses.  www.plycon.com  is a good source for heatsinks.  Not sure if they service the U.K. though.  
As for the voltage, do what Bombjack says.  The smallest increments you can until you get stability.  Usually a .1 V jump is all you need.  Beyond .1V the heat starts to really build and you run the risk of internal arcing in the chips core.  You can burn chips up playing with the voltage so be warned.  However I've got a K6-2 350 running at 450mhz after I bumped the voltage from 2.0 to 2.8 and its been alive in my office computer for over 2 years now.  But you should consider .1V your limit on a PIII.

[Gunthr]

Lynx,If you are running a BX6-R2, you can go up to a P-3 700 mhz cpu. I would do that if you can afford it. I ordered one from Googlegear a couple weeks ago for $243 US dollars for the boxed version. Although not advertised as such, it was a cbo stepping. Right out of the box with stock fan/heatsink it went up to 933 mhz just by upping the FSB to 133 mhz with 133mhz ram.

If not a 700, I would go for the 650 over the 600... you will get that much more out of it, all things being equal, and its less in demand, and maybe even the same price.

Cheers,


     

http://www.ropescourse.org/flying.htm"" TARGET=_blank>332nd Flying Mongrels




[This message has been edited by Gunthr (edited 06-29-2000).]

[NUTTZ]

Like i said , i didn't want to confuse about talking apples and oranges, I DO NOT have the coppermine series. If i miss quoted and called my chip the EB series then i was mistaken. I do have a pentium 600 that has 512 cache on the chip and running 133 bus See pricewatch.com for the price and availability of the chip. My momboard is the Asus P3V4X board pretty reliable. ( I also put the dual fans on the chip, Ya never can be to careful) As far as my over abundance of Ram.... I do graphics all day long, that is MY reasoning for a fast , yet very stable platform.

NUTTZ