| Author |
Message |
Guest
|
 Posted:
Tue Feb 01, 2005 12:23 am Post subject:
Re: History of performance counters |
|
|
Lee Witten <lw99@yahoo.com> writes:
| Quote: | I was at DEC during the early Alpha days (1992 or so), and they had
performance counters on the first Alpha chip and all the follow ons.
I recall we documented how they worked via the usual programming
references i.e. man pages for uprofile and kprofile on OSF/1 and
documents on IPROBE for VMS. But I don't think we particularly
documented how a third party could use them, although I think there
was obscure information on the kprof system call in the OSF/1 man
pages.
|
The perf counters where all described in the chip data `sheets'/HRMs.
| Quote: | I recall reading there may have been some profile counter support on
the first pentium (p5) and there definitely were on the second (p6).
So, I'm wondering if there are any earlier implementations of
performance counters, and when they were publically described. Were
the counters generally available to third parties, or where they
only really usable via vendor-provided tools. Who was the first to
provide the 'interrupt every N events of interest' style of
performance counters? Are there any relevant papers, patents, etc?
|
KL10s had them. Perhaps KIs as well. If you count adding jumper wires
to the backplane, the KB-11s had A counter that you could set up to
count how many time you hit a program set micro-address. And a few
other things... All of them where undoced folklaw.
--
Paul Repacholi 1 Crescent Rd.,
+61 (08) 9257-1001 Kalamunda.
West Australia 6076
comp.os.vms,- The Older, Grumpier Slashdot
Raw, Cooked or Well-done, it's all half baked.
EPIC, The Architecture of the future, always has been, always will be. |
|
| Back to top |
|
 |
Charlie Gibbs
Guest
|
| Posted:
Tue Feb 01, 2005 4:08 am Post subject:
Re: History of performance counters |
|
|
In article <MPG.1c6822a4890f932f9898c5@news.individual.net>,
krw@att.bizzzz (K Williams) writes:
| Quote: | Dunno, I've seen "designed for Microsoft Windows" stickers on urinals.
|
And you got your shoes wet too, I'll bet. Sounds like a GPF
(General Piss Fault). Did the water in the urinal turn blue?
--
/~\ cgibbs@kltpzyxm.invalid (Charlie Gibbs)
\ / I'm really at ac.dekanfrus if you read it the right way.
X Top-posted messages will probably be ignored. See RFC1855.
/ \ HTML will DEFINITELY be ignored. Join the ASCII ribbon campaign! |
|
| Back to top |
|
|
Brian Inglis
Guest
|
| Posted:
Wed Feb 02, 2005 7:56 am Post subject:
Re: History of performance counters (was: New Linux Power5 e |
|
|
On 29 Jan 2005 19:43:37 GMT in alt.folklore.computers, Lee Witten
<lw99@yahoo.com> wrote:
| Quote: | This posting made me wonder about the history of performance counters in
general, and the history of providing public information of their use.
I was an IBM employee in the late 80s / early 90s. At one point, I was
working on UNIX for the 3090 mainframes, and I saw a guy carrying a
listing several inches thick down the hall. I asked a co-worker who he
was and what he was doing, and he said he was from IBM Research, and the
listing was a dump of performance counters gathered when the kernel was
running, and he was analyzing them. I asked how I could do this, and
was told that was confidential info, and only certain priviledged folks
could learn how.
|
IBM 3x0 tracing and program event recording facilities were available
and documented as VM CP commands.
| Quote: | So, I'm wondering if there are any earlier implementations of
performance counters, and when they were publically described. Were the
counters generally available to third parties, or where they only really
usable via vendor-provided tools. Who was the first to provide the
'interrupt every N events of interest' style of performance counters?
Are there any relevant papers, patents, etc?
|
IBM 3x0 provided tracing and program event recording in at least S/370
hardware, perhaps earlier. For recent version see:
http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/DZ9AR004/4.2
thru 4.5...
--
Thanks. Take care, Brian Inglis Calgary, Alberta, Canada
Brian.Inglis@CSi.com (Brian[dot]Inglis{at}SystematicSW[dot]ab[dot]ca)
fake address use address above to reply |
|
| Back to top |
|
|
Alex Gibson
Guest
|
| Posted:
Sat Feb 05, 2005 9:58 pm Post subject:
Re: History of performance counters (was: New Linux Power5 e |
|
|
"Lee Witten" <lw99@yahoo.com> wrote in message
news:Xns95ED95CEF5AF9nn48@199.125.85.9...
| Quote: | anton@mips.complang.tuwien.ac.at (Anton Ertl) wrote in
news:2005Jan25.091041@mips.complang.tuwien.ac.at:
An advantage of the PPC74xx over the PPC970 is that information about
the performance monitoring counters is publically available, and
therefore the Linux perfctr patch supports the PPC74xx, whereas IBM
has not provided that information for the PPC970, and therefore the
perfctr patch does not support the PPC970. That's fairly important
for me, and I would expect it to be even more important for Terje.
(For more information on how to use perfctr's perfex on the PPC7450
familiy, see
http://www.complang.tuwien.ac.at/anton/linux-perfex/ppc7450.html>).
This posting made me wonder about the history of performance counters in
general, and the history of providing public information of their use.
I was an IBM employee in the late 80s / early 90s. At one point, I was
working on UNIX for the 3090 mainframes, and I saw a guy carrying a
listing several inches thick down the hall. I asked a co-worker who he
was and what he was doing, and he said he was from IBM Research, and the
listing was a dump of performance counters gathered when the kernel was
running, and he was analyzing them. I asked how I could do this, and
was told that was confidential info, and only certain priviledged folks
could learn how.
At a similar point in time, I was logged on to a certain system within
IBM, and I read that the first generation RS/6000s had some sort of
performance counters that could only be read if you had a specially
modified motherboard with special probes hooked up to read them. I
never found out any more about them.
I was at DEC during the early Alpha days (1992 or so), and they had
performance counters on the first Alpha chip and all the follow ons. I
recall we documented how they worked via the usual programming
references i.e. man pages for uprofile and kprofile on OSF/1 and
documents on IPROBE for VMS. But I don't think we particularly
documented how a third party could use them, although I think there was
obscure information on the kprof system call in the OSF/1 man pages.
I recall reading there may have been some profile counter support on the
first pentium (p5) and there definitely were on the second (p6).
So, I'm wondering if there are any earlier implementations of
performance counters, and when they were publically described. Were the
counters generally available to third parties, or where they only really
usable via vendor-provided tools. Who was the first to provide the
'interrupt every N events of interest' style of performance counters?
Are there any relevant papers, patents, etc?
--lw--
|
I believe Apple's shark debugger uses the performance counters
in the 970 / G5 |
|
| Back to top |
|
|
Anne & Lynn Wheeler
Guest
|
| Posted:
Sat Feb 05, 2005 10:30 pm Post subject:
Re: History of performance counters |
|
|
"del cecchi" <dcecchi.nojunk@att.net> writes:
| Quote: | I think in Poughkeepsie the location of the rest rooms was IBM
Confidential. :-)
In Rochester, the idea that the CE would use a scope was totally
foreign. We were lucky if we could get them to use a voltmeter. There
was a "logic probe" switchable for MST or SLT that would light a light
to show 1 or 0 on a pin.
|
the following is a tail of online telephone book application
http://www.garlic.com/~lynn/2005c.html#38 [Lit.] Buffer overruns
one of the most difficult challenges was convincing various site
security people that softcopy of the plant telephone directory didn't
have to be treated as confidential ... it could be considered simply
"internal use only" (as was the paper copy of the same).
once the security people at one plant site was convinced ... we could
use that fact in arguments with other plant site security people.
--
Anne & Lynn Wheeler | http://www.garlic.com/~lynn/ |
|
| Back to top |
|
|
Anne & Lynn Wheeler
Guest
|
| Posted:
Sat Feb 05, 2005 10:39 pm Post subject:
Re: History of performance counters |
|
|
Lee Witten <lw99@yahoo.com> writes:
| Quote: | When I worked there, every terminal and computer had a sticker that
said "This equipment can be used for IBM Mangement Approved Purposes
only", or some such statement. And I got a huge laugh when someone
put one of those stickers on all the urinals in the IBM Mens Room!
Such naughtyness was rare at IBM!
|
part of that was from an early incident involving corporate people
coming into doing a site security audit in san jose. we eventually
escalated with a challenge to the security people that people's
personal computing facilities should be treated similarly to their
physical office, desk, phone, etc .... and we provided a draft update
to some of the wording for corporate management manuals to reflect the
situation.
we also had some confrontations with them about the available of
"games" on the system. we asserted that it would be better to have
single copy of games in public area on the system ... then everybody
having their own individual disguised copy ... and furthermore they
were quite educations.
I had gotten a very early copy of Fortran source from Tymshare that
they had ported from stanford sail to their vm370/cms timesharing
service ... and I was making it available inside the company (I would
also send the source to anybody that could proove they made all
points). I was asserting that it was a good exposure to some more
interesting interactive human factors (than most of them had been
exposed to with online systems up until then). At one point, the
executives at STL/bldg.90 claimed that 1/3rd of all computer time
appeared to be consumed playing adventure ... and they gave notice of
24hr grace period and after that no 1st shift playing of adventure
would be tolerated.
--
Anne & Lynn Wheeler | http://www.garlic.com/~lynn/ |
|
| Back to top |
|
|
Anne & Lynn Wheeler
Guest
|
| Posted:
Sat Feb 05, 2005 10:55 pm Post subject:
Re: History of performance counters |
|
|
note, one of the wording changes was from "Business Use Only" ... we
pointed out that it was management discretion that people may do
things like talk to their spouse or children on their desk phone
during the course of the day ... it it didn't impact business
operations. The analogy that computers shouldn't be treated any
differently than any other business facility.
--
Anne & Lynn Wheeler | http://www.garlic.com/~lynn/ |
|
| Back to top |
|
|
Charles Shannon Hendrix
Guest
|
| Posted:
Sun Feb 13, 2005 1:09 am Post subject:
Re: History of performance counters (was: New Linux Power5 e |
|
|
["Followup-To:" header set to comp.arch.]
On 2005-01-29, Terje Mathisen <terje.mathisen@hda.hydro.com> wrote:
| Quote: | Lee Witten wrote:
I recall reading there may have been some profile counter support on the
first pentium (p5) and there definitely were on the second (p6).
So, I'm wondering if there are any earlier implementations of
performance counters, and when they were publically described. Were the
The Pentium counters were undocumented, except by NDA, so it took me a
year from I first heard about it until I had reverse engineered them and
documented their working for a Byte article in 1994.
|
That's interesting, because I remember running a debugger on Linux
around that time that gave you profile information kind of like gprof,
but was using the Pentium hardware counters.
It had a small utility to draw related graphs too.
I wonder if it was based on your article?
I'd like to do the same thing with my current AMD64 system. Have yet to
really look and see if it is even possible.
It might even be that modern Linux and *BSD kernels collect this
information by default now.
--
shannon "AT" widomaker.com -- ["There is no such thing as security. Life
is either bold adventure, or it is nothing -- Helen Keller"]
----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =---- |
|
| Back to top |
|
|
Scott A Crosby
Guest
|
| Posted:
Sun Feb 13, 2005 6:32 am Post subject:
Re: History of performance counters |
|
|
On Sat, 12 Feb 2005 15:09:17 -0500, Charles Shannon Hendrix <shannon@news.widomaker.com> writes:
| Quote: | That's interesting, because I remember running a debugger on Linux
around that time that gave you profile information kind of like gprof,
but was using the Pentium hardware counters.
I'd like to do the same thing with my current AMD64 system. Have yet to
really look and see if it is even possible.
It might even be that modern Linux and *BSD kernels collect this
information by default now.
|
Oprofile is a user/kernel full system profiler for linux 2.4.x and
linux 2.6.x kernels. It ties into the performance counters. I use it
constantly.
Scott |
|
| Back to top |
|
|
Eugene Miya
Guest
|
| Posted:
Thu Feb 17, 2005 1:23 pm Post subject:
Re: History of performance counters (was: New Linux Power5 e |
|
|
In article <3669miF4soejsU1@individual.net>,
=?ISO-8859-1?Q?Jan_Vorbr=FCggen?= <jvorbrueggen-not@mediasec.de> wrote:
| Quote: | I believe the Crays has a set of hardware performance counters, and that
the interface on using them way documented. Anybody remember details?
|
Oh yes, I wrote a paper about calibrating them using 3 different
machines/systems under four different languages. X-MPs had 8 registers
with 4 modes and Y-MP/C-90s skipped the modes and just gave 32 registers.
It's not a topic people consider much. In the X-MP COS days (and CTSS)
you accessed them via Fortran subroutine calls and after that in Unicos
days you used a command like time(1).
However after the paper i realize that there was a subtle measurement
inaccuracy. The loops following this comment:
/* zero out the holding and receiving
* areas for the counters
*/
sets those counters to 0 in a bad order (low numbers to high).
This should have been reversed. So since that's just a comment which
any similar hardware would have, I doubt this violates NDA.
Anyways, the paper got the Knuth nod (he really liked it, my management
thought it was trivial), and then I got the invitation to give the paper at Fort
Meade. Quite amusing.
A lot of non determinism in those counters.
-- |
|
| Back to top |
|
|
Grumble
Guest
|
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in