View Full Version : PowerPC 750fx - very slow problem

25th September 2003, 14:09
What would be the most likely mistakes to make in programming a PowerPC 750fx chip to make it run like Tiny Tim?

25th September 2003, 14:46
Data Alignment....?

25th September 2003, 20:49
> What would be the most likely mistakes to make in
> programming a PowerPC 750fx chip

making a fundamentally wrong decision to support failed chip - its not 1983 it is 2003! :lol

26th September 2003, 08:50
Shows what you know you commie twat!

IBM - PowerPC 750FX (Anaconda)
The 750FX is produced on a .13 micron process with Low-K Dielectric and Silicon-on-Insulator technology. There is greatly improved cache performance as well with a 512 KB L2 cache (opposed to 256 KB of cache in earlier 750CXe). There's also 256-bit cache lines instead of 64-bit, in addition to improved logic. That, along with the 200 MHz bus speed improve performance by 25% over the 750CXe at the same clock speed.

26th September 2003, 11:40
Bit faster than a 6800 then...

26th September 2003, 21:55
> Shows what you know you commie twat!

thanks mate :D

What I know is that x86 cpus give more bang for the buck - smart people like Google know what they doing by using lots of cheap boxes with x86 cpus. As far as I am concerned Motorola's designs died the day Intel reached 90% market share.

28th September 2003, 10:28
Another one of those idiotic arguments. Having a small market share doesn't mean you don't have a hugely profitable company.

For example, Apple only sell a small percentage of the total number of PCs, however, they have been/are/were in the top five PC manufacturers in the world. Not something to be sniffed at.

We all know that having the major share of the market is very rarely due to design or technical excellence.

28th September 2003, 14:02

Different horses for different courses. I notice you were being your usual helpful self!

Threaded, have you checked your data alignment?

28th September 2003, 20:46
OrangeHopper: agree with you but not in regards to Motorola and Apple - the former is struggling to become profitable and the latter's profitability is not that great.

> however, they have been/are/were in the top five PC
> manufacturers in the world.

this is a misleading figure - a LOT of people buy their computers from local small companies, so martket below Top 10 is extremely fragmented but overall these small companies have more than 50% of sales (last time I checked).

> We all know that having the major share of the market is
> very rarely due to design or technical excellence.

Thank God that "design or technical excellence" do not win all the time - sorry but there are practical things in this world such as cost of processor - Intel won on cost and compatibility grounds.

Just wait till Apple switches to x86 architecture in 3-5 years time. Remember - you heard it here first!

Anyway, since I am not helping as Spod correctly pointed out, I will shut up.

1st October 2003, 13:54
Spod: that was an additional issue, but the main one was something else really dumb and the client won't thank me for telling you all they've spent several months looking for a hardware mistake in the software... Doh!

1st October 2003, 16:18
Don't you just love that, specially if you're the one who designed the hardware (speaking from experience here).

My favorite was a 64 pin device that ended up with only 63 pins on the pcb. Ooooops.

6th October 2003, 20:57
yadda yadad yadda SupremeSpod - read this about your belowed Motorola PowerPC - "Motorola to spin off chip unit"

news.com.com/2100-1006-5086763.html (http://news.com.com/2100-1006-5086763.html)

do you know what it means? Yes, this is beginning of the end! IBM will still use it for servers but watch Apple switching to AMD or even Intel. Jobs made smart move by using portable BSD Unix.

7th October 2003, 08:37
So what, Motorola spun off parts of the semiconductor operation some years ago.

7th October 2003, 20:26
this time PowerPC goes - Motorola is focusing on other things, and this will put extra pressure on Apple to move to x86 architecture. It is not easy thing to do for Jobs but he is a hard man and can do it.

8th October 2003, 11:26
IBM will still use it for servers

And Cisco will still use it for their comms boxes, and Mitel for their switches, along with hundreds of other embedded users. What is your point?

8th October 2003, 12:14
Quote from some magazine "Embedded: the refuge of the failed desktop processor".

8th October 2003, 20:32
my point is that by disappearing from Personal COmptuers and going fully emebedded Motorola ends the era proving the point that in its time Intel made right decisions cost, CISC, backwards compatibility wise.

9th October 2003, 08:47
I'm sure that those of us who programmed the 8080 would agree...

9th October 2003, 19:33
dont knock intel - they opened new era of cheap general purpose microprocessors by releasing 4004, and 8080 was important step towards 8086.

10th October 2003, 08:46
ATW, I remember it well... (through the mists of time, there stands before me the Intel Blue Box MDS with elephant's trunk processor interface).

And we thought 64k was a lot after the 16k of the 8008.

10th October 2003, 20:34
the point is that x86 has won it - test of time shown that the choices and sacrifices that Intel made 25 years ago paid back very nicely. Mainstream software compatibility is valued by consumers far more that nice looks of Apple - I respect Jobs a lot and I wish I was like him but with all that due respect I am damn grateful he did not lock the whole world into his closed way of thinking - remember, they have monopoly on making Apple Computers, something that doomed him to niche market.

11th October 2003, 08:44
But then again, Apple are still making Macs...


11th October 2003, 11:42
the only companies that really make moneyin x86 space is Intel + add-on manufacturers (gfx cards, hard drives), but PC asselmlers which is what Apples does - are working on razor thin margins. Who wants to be in that business? IBM does not and i cant blame them.

11th October 2003, 22:35
Somewhere or other I have a utility for converting 8080 code to 8086 code.

Dread to think what the resulting code looks like though.

IBM shot themselves in the foot with MicroChannel and the PS series.

12th October 2003, 14:02
true microchannel was a failure, but more monumental failure of IBM was to recognize that low cost producers like Compaq will have their day. It was real shame to see Compaq go as it was them who freed BIOS and allowed low cost production of PCs.