Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Thu, 12 Sep 2013 19:23:35 +0200
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: PUTCHAR() macro

On 12 sep 2013, at 15:23, Sayantan Datta <std2048@...il.com> wrote:
> On Thu, Sep 12, 2013 at 12:27 PM, magnum <john.magnum@...hmail.com> wrote:
>> I'd think that macro would need its share of registers too. What's the quickest/simplest way to see vgpr use and other things for AMD, without using any GUI? Is this on wiki somewhere?
> 
> Use -save-temps in build option. It would generate *.isa file which contains all the details near the end of file.

Thanks. The vgpr count actually decreases a lot, but scratchsize increases:

bit-flogging to 32-bit stores:
userElementCount     = 3;
;  userElements[0]      = PTR_UAV_TABLE, Element=1 dwords, s[2:3]
;  userElements[1]      = IMM_CONST_BUFFER 0, s[4:7]
;  userElements[2]      = IMM_CONST_BUFFER 1, s[8:11]
extUserElementCount  = 0;
NumVgprs             = 115;
NumSgprs             = 46;
FloatMode            = 192;
IeeeMode             = 0;
ScratchSize          = 0;


byte-addressed stores:
userElementCount     = 4;
;  userElements[0]      = PTR_INTERNAL_GLOBAL_TABLE, Element=1 dwords, s[0:1]
;  userElements[1]      = PTR_UAV_TABLE, Element=1 dwords, s[2:3]
;  userElements[2]      = IMM_CONST_BUFFER 0, s[4:7]
;  userElements[3]      = IMM_CONST_BUFFER 1, s[8:11]
extUserElementCount  = 0;
NumVgprs             = 34;
NumSgprs             = 38;
FloatMode            = 192;
IeeeMode             = 0;
ScratchSize          = 72;

I guess it starts using slow memory as scratch space.

magnum

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ