gamedevlists-windows

[GD-Windows] profiler for AMD

[Tugkan C. <tu...@in...>]

We are considering to switch to Athlon XP from Intel chips.We know that =
XP is compatible with
Intel instruction set but  are they compatible with intel architecture =
to the extent of internal counters?
What I am after is, indeed, whether I could use VTune when I use AMD =
chip?
If not, is AMD's CodeAnalyst good enough to meet development needs ( =
currently it has version 1.1 which seems like
they did'nt have a long history of making profilers :) ) ?=20

Re: [GD-Windows] profiler for AMD

[Neil S. <ne...@r0...>]

On pre-XP Athlons, you couldn't use Event Based Sampling (EBS) in VTune,
which uses Intel's internal counters, but you could still use Time Based
Sampling (TBS), using various timers (eg. Virtual Timer Device), which
should be good enough for most purposes. I don't think this has changed on
Athlon XP.

----- Original Message -----
From: "Tugkan Calapoglu" <tu...@in...>
To: <gam...@li...>
Sent: Wednesday, January 16, 2002 4:33 PM
Subject: [GD-Windows] profiler for AMD


We are considering to switch to Athlon XP from Intel chips.We know that XP
is compatible with
Intel instruction set but  are they compatible with intel architecture to
the extent of internal counters?
What I am after is, indeed, whether I could use VTune when I use AMD chip?
If not, is AMD's CodeAnalyst good enough to meet development needs
 currently it has version 1.1 which seems like
they did'nt have a long history of making profilers :) ) ?

[GD-Windows] Spin Control Right Align "Upside Down"

[Corrinne Y. <cor...@sp...>]

This is for the editor of my engine.

I have an auto-buddy-ed Spin Control that is aligned right to be edit
box.

It is such that artists can type in number of, use the spin to scroll it
bigger or smaller.

I notice that when you click the "up arrow" the numbers get bigger, and
the "down arrow" the numbers get smaller.

This seems upside down (or backwards) to me. :)

I figure in order to get the direction I want, I have to not use the
convenient auto-buddy, and manually link the edit box number and the
spin control value.

Before I waste time writing that code, I wonder if you know of a special
flag that would make the spin works "upside down" than from before (or a
way to call the Windows API to do it for you).

Thank you for any info from you Windows Experts out there. :)

RE: [GD-Windows] profiler for AMD

[Jon W. <hp...@mi...>]

VTune does not work very well with AMD processors; it degrades to
timer-based sampling (yeah, the 8253 lives! :-)

CodeAnalyst is where it's at for AMD -- it's certainly better than
nothing, but I haven't used it enough to say more (i e sniffed it
once)

Even though VTune is at a high version number, it has its annoying
habits, such as blue-screening Win2k now and then, and being very
slow at parsing symbol files. I believe most of it is written in
Visual Basic (except for the device driver part). If you think VTune
is doing fine, I wouldn't be scared of trying CodeAnalyst.

Cheers,

			/ h+


-----Original Message-----
From: gam...@li...
[mailto:gam...@li...]On Behalf Of Tugkan
Calapoglu
Sent: Wednesday, January 16, 2002 8:33 AM
To: gam...@li...
Subject: [GD-Windows] profiler for AMD


We are considering to switch to Athlon XP from Intel chips.We know that XP
is compatible with
Intel instruction set but  are they compatible with intel architecture to
the extent of internal counters?
What I am after is, indeed, whether I could use VTune when I use AMD chip?
If not, is AMD's CodeAnalyst good enough to meet development needs (
currently it has version 1.1 which seems like
they did'nt have a long history of making profilers :) ) ?

Re: [GD-Windows] profiler for AMD

[Eero P. <epa...@ko...>]

Tugkan Calapoglu wrote:

 > We are considering to switch to Athlon XP from Intel chips.We know that
 > XP is compatible with
 >
 > Intel instruction set but  are they compatible with intel architecture
 > to the extent of internal counters?
 >
 > What I am after is, indeed, whether I could use VTune when I use AMD chip?
 >
 > If not, is AMD's CodeAnalyst good enough to meet development needs (
 > currently it has version 1.1 which seems like
 >
 > they did'nt have a long history of making profilers :) ) ?
 >

Considering how often Vtune (I have only tested it up to
version 4.5) decides that all my CPU is used by "other32",
I have been rather happy with AMD codeAnalyst.

You can also use VTune with Athlon, except for the EBS based sampling.



		Eero

Re: [GD-Windows] profiler for AMD

[Tugkan C. <tu...@in...>]

----- Original Message -----
From: "Eero Pajarre" <epa...@ko...>
To: "Tugkan Calapoglu" <tu...@in...>
Cc: <gam...@li...>
Sent: Thursday, January 17, 2002 3:21 PM
Subject: Re: [GD-Windows] profiler for AMD


> Tugkan Calapoglu wrote:
>
>  > We are considering to switch to Athlon XP from Intel chips.We know that
>  > XP is compatible with
>  >
>  > Intel instruction set but  are they compatible with intel architecture
>  > to the extent of internal counters?
>  >
>  > What I am after is, indeed, whether I could use VTune when I use AMD
chip?
>  >
>  > If not, is AMD's CodeAnalyst good enough to meet development needs (
>  > currently it has version 1.1 which seems like
>  >
>  > they did'nt have a long history of making profilers :) ) ?
>  >
>
> Considering how often Vtune (I have only tested it up to
> version 4.5) decides that all my CPU is used by "other32",
> I have been rather happy with AMD codeAnalyst.
>
> You can also use VTune with Athlon, except for the EBS based sampling.
>
>
>
> Eero


That 'other32' and 'library' is really a problem for me ( Some bars say they
are library.There is no clue about
which library they are ) I am not sure whether it is problem of VTune or
compiler. I think compiler could not succeed to
insert debug info or VTune has problems in interpreting it. I worked around
it by selecting "line numbers only" option.
My question to Intel's support service remained unanswered ( Well they gave
an answer which is totally useless. )
But if CodeAnalyst does a better job than VTune in this subject, this is a
real breakthrough  :).