[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: __fres() pipeline depth

Subject: Re: __fres() pipeline depth
From: Ian Ollmann <email@hidden>
Date: Wed, 19 Apr 2006 11:35:40 -0700
Delivered-to: email@hidden
Delivered-to: email@hidden


On Apr 19, 2006, at 9:35 AM, Sanjay Patel wrote:

I don't have a G3 to confirm...if someone does, please try the following test program:


I find my 700 MHz G3 to loop 700 million times in the following time:

no loop:	0.002s
1 nop:	2.0s
1 fres:	4.0s
2 fres:	4.0s
3 fres:	6.0s
4 fres:	8.0s
5 fres:	13.0s

If the fres are dependent on one another (that is, we are measuring latency) then I get 4 seconds per fres in the loop. I have no explanation for the oddball result at 5 fres.

Using ppc_intrinsics.h on gcc3.3 causes some extra fmr to be issued, so you need to write your own asms to get this to compile as intended.

Ian
_______________________________________________
Do not post admin requests to the list. They will be ignored.
PerfOptimization-dev mailing list      (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/perfoptimization-dev/email@hidden

This email sent to email@hidden

References:
	>Re: __fres() pipeline depth (From: Sanjay Patel <email@hidden>)

Prev by Date: problem with FireStarter installation
Next by Date: Re: XCode 2.2.1 / gcc 4.0 Peephole Bug
Previous by thread: Re: __fres() pipeline depth
Next by thread: Saturn?
Index(es):
- Date
- Thread

Home