CPU benchmarking - this sucks

  • Thread starter Thread starter Flasherly
  • Start date Start date
F

Flasherly

Pulled out the card table. I got 4 arms, btw, for any intents. And
they're in split-second synch.

Yes, I can run two keyboards, mice, and computers, both at once (don't
try this in your home, though.)

1system is P4 3Ghz single core

2system is P4-E 2.66 dual core

Benchmarks aren't cutting it - so...

I take out two instances of normalization for 1/2G of raw audio
performances - MP3 Gain program (on Rory Gallagher, a blues
guitarist.)

That makes 7 performance MP3 files split over both program - or on two
instances of MP3 Gain, sic., running twice. In the redundant sense of
iteracy

Also ran WTaskMngr for Performance/CPU hist - where both graphs aren't
telling me squat.

I run test results twice. Both instances, the 7 files on system2 dual
core are normalized faster. In fact, it's kicking little Bro's P4
3Ghz butt big time.

Couldn't somebody have just written a program that says that it that
simply? (Or maybe that wouldn't be covered under relativity rules.)
 
Also ran WTaskMngr for Performance/CPU hist - where both graphs aren't telling me squat.
Click to expand...

...s'xuse me. The SYS 1, P4 3Ghz single core is quasi-ated by graphic
intents into a dualcore due to wondrous HyperThreading. At least in
XP SP1 linguistics, which is all you really need to know,
i.e., ...They work, look good and gracious while displaying two cores
for inimically pretty microprocessor clones, at all times shown highly
green at maxed-out warp speed under loads, under disparate affinity
assignments over disparate files. I could update for a newer and
approved SP, learn to read graphic graphs better, or best yet, I
suppose, run time tests without it. (Didn't somebody say I can't do
this on SP1? I remember it was hard to get it on, cloned and
crashing, bitting and gnawing, and screaming all the way across my old
single to a dual platforms, but I did and then promptly forgot how.
That how it gets grandfathered in now.)
 
Flasherly said:
Pulled out the card table. I got 4 arms, btw, for any intents. And
they're in split-second synch.

Yes, I can run two keyboards, mice, and computers, both at once (don't
try this in your home, though.)

1system is P4 3Ghz single core

2system is P4-E 2.66 dual core

Benchmarks aren't cutting it - so...

I take out two instances of normalization for 1/2G of raw audio
performances - MP3 Gain program (on Rory Gallagher, a blues
guitarist.)

That makes 7 performance MP3 files split over both program - or on two
instances of MP3 Gain, sic., running twice. In the redundant sense of
iteracy

Also ran WTaskMngr for Performance/CPU hist - where both graphs aren't
telling me squat.

I run test results twice. Both instances, the 7 files on system2 dual
core are normalized faster. In fact, it's kicking little Bro's P4
3Ghz butt big time.

Couldn't somebody have just written a program that says that it that
simply? (Or maybe that wouldn't be covered under relativity rules.)
Click to expand...

Ignoring the effects of the Hyperthreading, 2*2.66 > 1*3.0.

And this only holds true, if a program uses both cores.
Not all audio programs are multi-threaded that way.
You could easily end up with 1*2.66 < 1*3.0, if the
software isn't multi-threaded.

If you need to rate them, look in the list.

http://www.cpubenchmark.net/cpu_list.php

Paul
 
Ignoring the effects of the Hyperthreading, 2*2.66 > 1*3.0.

And this only holds true, if a program uses both cores.
Not all audio programs are multi-threaded that way.
You could easily end up with 1*2.66 < 1*3.0, if the
software isn't multi-threaded.

If you need to rate them, look in the list.
Click to expand...

I need to determine if it's worth is commiserate with effort to
continue along and swap out this 3Ghz system with that 2.6x2CPU.
Optimistically, the plan (the 2x2.66 does have a full XP's 2G of
memory, this has 1G);- the SSD in it, I could swap to here if tests
were a big letdown. Just a care and thought first to back-checking
performance wise, IOW. All that audio normalization stuff is only a
program that runs over a MP3 file for analysis;- heats up the CPU,
then, determines volume peaks over an entirety and resets the MP3's
imbedded header, perhaps from the recording studio, to an alternative
and user-preference as it's end and aim. Big plus and affirmative.
I'm not getting that performance on the 3Ghz when running two
instances if similarly processing multiple instances of audio
files.... 'Ignoring the hyperthreading.' Yes, I do realize that
snafus exist, (one may questionably be SP1's unawareness/drive core
issues for a dualcore, however I managed that, to shoe-horn this OS
onto that platform, a couple months ago), that, once past and
apparently running without further issues, tweaking for affinity and
such is well within a benevolent efficiency design;- the realization
there's advantage to the 2x2.66Ghz in this one instance (from the
programs related and tested), well, that's simply a step in the right
direction past this 3Ghz. It can go into the backroom and porch now
for parts, a backup, or someone's interest and hard-earned money for
something else I no doubt deserve. :)
 
Back
Top