The case of the 6500 and the missing cache

[eraser]

Short version: My PowerMac 6500/275 sees 256K of L2 Cache although a 512K module is installed.

Long version:

I acquired my PowerMac 6500 some years ago and it has been an awesome machine. Although it is a 275 MHz version it only had a 256K cache module. I'm not sure why, but I wasn't the first owner so I figured maybe the previous owner had used one of the G3 upgrades and put the wrong module back in. Well, I found a Mac-friendly memory retailer who happened to have a 512K module for sale. I bought it.

Now, the 512K module I received has the same pin number and spacing as the 256K module installed but, as expected, is a bit taller and much more densely populated. There are 16 memory chips labelled IDT 71256 (32K) and 16 x 32K = 512K. This does appear to be a 512K module. When I install the module though, Apple System Profiler says 256K external L2 cache.

Any one know why?

[waynestewart]

Don't know if this'll help but my 6500 512k cache have stickers with the part number IDT7MPV6284, while my 256k modules have the part number IDT7MPV6283

[Byrd]

I've found Apple System Profiler to often be wrong detecting cache size. Being unsure of another application to check it, the only thing would be to benchmark the modules (eg. Macbench, Norton System Info), but be aware the performance difference will only be very slight

[BGoins12]

My 5500/250 with a 6500/275 board does the same exact thing. It shows the cache as 256k, yet the part number shows it as a 512k module.

[eraser]

Don't know if this'll help but my 6500 512k cache have stickers with the part number IDT7MPV6284, while my 256k modules have the part number IDT7MPV6283

Are these numbers reversed? I have both of these modules and it looks like the opposite is the case.

I've found Apple System Profiler to often be wrong detecting cache size. Being unsure of another application to check it, the only thing would be to benchmark the modules (eg. Macbench, Norton System Info), but be aware the performance difference will only be very slight

Ah! Is there a tool that is more reliable (Guage Pro?)?

In that case I did see a very small difference in the Norton benchmark but I was surprised at how slight it was. I kindof figured it was a change in some extension or display setting vs. the difference in cache. Wow, I guess doubling the cache doesn't make all that much difference.

Thanks for the help guys!

[Byrd]

On machines of this era, adding more cache is a law of diminishing returns. But if you want to squeeze as much performance out of your machine as possible, it is a cheap upgrade. The performance difference won't be noticeable in day-to-day use, but every little helps

[eraser]

On machines of this era, adding more cache is a law of diminishing returns. But if you want to squeeze as much performance out of your machine as possible, it is a cheap upgrade.

Indeed. I am aware of the diminishing returns when it comes to cache although I was a little surprised at how diminished it was.

Then again in the Norton benchmark the actual app is small and I suspect that much of the test code fit in the 256K cache and that going to 512K didn't show up in the benchmarks due to this. I will have to experiment a bit and see if there is an improvement in 'real world' cases.

Thanks again

[Trash80toHP_Mini]

The next step up for the 6500 would 1MB of cache on a Sonnet G3 Card in that slot . . . very little noticeable in the way of diminishing returns!

[gobabushka]

when i put my g3 card in my c500, I noticed a large difference in speed. but i also went from 200Mhz 603e to 500MHz G3. I never did benchmark it tho...

[waynestewart]

I checked and those numbers are correct. I have two of each and they show up as those sizes in both System Profiler and GuagesPro.

Are these numbers reversed? I have both of these modules and it looks like the opposite is the case.