How do I set latency slider in ASIO mode to determine latencies? My Echo Mia card only allows buffer changes in ASIO. I maybe doing something incorrect, not sure.

You don't actually move the slider. Click the "ASIO Panel" button in Audio Options and set the buffer size in the Mia Control Panel.

-S

Well at first Sonar wouldn't load the plugs. So I installed the 3.1 patch and... it worked. I'll add another Athlon score.

System Specs
Motherboard: Asus A7N8X deluxe (nForce2 dual channel)
RAM: 512Mb Corsir TwinX Timed at 2-2-2-5
Processor: AthlonXP 2700+ T-Bred B core
FSB: 333 MHz
CPU: 2.17 GHz
Multiplier: 13
Audio Device: WDM M-Audio Delta 0.29 driver set to 64 sample buffer

46.4ms: 33%
23.2ms: 33-36%
11.6ms: 39-43%
5.8ms: 51-58%
2.9ms: 73-76%
1.5ms: Warning shown but doesn't drop out

My score observations so far (warning: contains speculation):

This benchmark shows the P4 in a much better light than the old sonarbench. This is most likely due to better optimizations for SSE and other compiler related improvements for the P4. The older plug-ins were most likely primarily x87 in nature and ran well with the Athlon FPU.

This particular test file seems to like memory bandwidth a whole lot. The 1GHz FSB P4 at 3.0GHz (timboe) turned out a better score than the 3.0GHz P4 with a stock 800MHz FSB (cAPSLOCK). The Athlon FX and Opteron scores are very high. This could be due to the higher bandwidth and lower latency memory access of those chips, but differences in architecture make it difficult to speculate. The Athlon XP chips fall approximately in line with the P4 equivalent. For example my XP2700+ scores curve very closely with the P4 2.6GHz with an 800Mhz FSB scores turned in by mistergarner, coinciding at 2.9ms and 23.2ms.
(HT was only enabled on the C version of the P4 so that Dell machine should have an 800MHz FSB)

With the current prices, that makes the sub $200 Athlon XP and the Pentium 4 very close in price vs. performance. The top of the line is also well matched, the $408 Opteron 146 should outperform the $385 P4 3.2GHz by about that much money. The $700 Athlon 64 FX 51 is definitely the single processor champion, but I can’t say it’s worth the extra cash (sorry Scott, but I’m still jealous). I would have to guess that the 2MB L3 cache of the P4 Extreme Edition isn’t going to make enough of a difference in Sonar to make it worth $915.

The only puzzle piece missing is the standard Athlon 64 chips, as mentioned by Niko. That may indicate if Sonar responds well to the improved memory latency, the bandwidth, or most likely, a combination of the two. If the Athlon 64s perform as well as their big brothers, they would be the price vs. performance champion.
(Prices from www.pricewatch.com)

if i felt industrious today i'd load all of this data into a database & make it
searchable online.

but instead i gues i should do other work.

Something is out of whack on that 2800MP system

ORIGINAL: Scott Reams
So:

The single-CPU king of the hill would be the $404 Athlon64 3400+ (2.2GHz, 1MB L2) in this test. The best value would most likely go to the Athlon64 3000+ (2GHz, 512kb L2, $211) or Athlon64 3200+ (2GHz, 1MB L2, $270). At least so far. Of course... I suspect that dual Opterons will be the ultimate performance champion in this case.

How fast of a computer will I need to make *GOOD* music?

Also, will duel Opera-Trons make my hihat sound any better?

ORIGINAL: wogg
[snip]
The only puzzle piece missing is the standard Athlon 64 chips, as mentioned by Niko.

Well, the other missing piece would be a couple of Centrino/Pentium-M
scores. On the old Sonartest, and on the Reaktor benchmarks, they
were amazingly efficient.

ORIGINAL: Glennbo

How fast of a computer will I need to make *GOOD* music?

Also, will duel Opera-Trons make my hihat sound any better?

You are starting to sound a lot like Ted... :)

I know you are kidding around, but I hope you aren't implying that this information is not useful. We all know quite well that a DAW's processing capabilities are only one aspect of a larger equation.

Will a faster DAW make my music writing better? No. Will it smooth out the recording process... allowing more focus to be put into creativity? Sure. The less bouncing I'm doing, the more time I am spending on the music.

-S

ORIGINAL: Duojet

Something is out of whack on that 2800MP system

It looks about right to me. Some things to remember:

The 2800+ MP is a Barton CPU... and has a slower core clock than the 2700+ and 2600+ XP (2.08MHz vs 2.17/2.13MHz). It has a slower FSB than any of the more modern XP CPUs (266MHz vs 333 or 400MHz). It also has an older chipset by comparison. Newer single-CPU XP chipsets are better optimized overall.

Edit: Actually... looking closer, the dual 2800+ system isn't doing much better than the dual 2000+ system... despite a 25% clock speed advantage and twice the cache. It definitely deserves a second look.

-S

yes, but it looks like it should be further ahead of the 2000+ MP system. also, the 2700 xp system seems TOO FAR ahead(like you said, FSB etc). nevertheless this analysis is priceless. you couldnt find this anywhere else on the internet. most sites do some synthetic benchmarks and call it a day.

i thought the extra cache would have helped more than clock speed. also its interesting to see that the dual channel memory of the fx51 doesnt help much here. i think my next system will be a 3400 athlon 64.

the dual 2800+ system isn't doing much better than the dual 2000+ system

Great. Well there's a few hundred bucks and a long weekend down the toilet!

(just upgraded from 1900+'s)

Steven

ORIGINAL: Steven Bell

the dual 2800+ system isn't doing much better than the dual 2000+ system

Great. Well there's a few hundred bucks and a long weekend down the toilet!

(just upgraded from 1900+'s)

Steven

It's probably worth looking at why this is happening. Your system has every reason to be 20-25% faster than that dual 2000+.

-S

ORIGINAL: woody24

Boy....I'm getting lousy results!!
Anyone know why?

P4 1.8 Titan GA-8IE800 intel845E Chipset
512 DDR Ram
2-7200rpm HDrives
Delta66/Omni 27drivers

WinXP
Sonar3.1
Mackie Control Surface

WDM
46.4 65cpu
23.2 65cpu
11.6 78cpu
5.8 xxx
2.9 xxx
1.5 xxx

Anything I can do to lower my cpu ? Woody

Well with ASIO drivers I get about 10% improvement on all settings....

Should I be using ASIO instead of WDM?
I thought WDM were better???

And thanks a lot for taking the time to do all of this!
So my 1.8 is already to slow.....geeesh....I've only had it about 4 months!
Woody

ORIGINAL: Duojet

just a suggestion, might a line graph work better?

I just tried it...

It's better in one way... as it is easier to follow one CPUs performance across several latencies. It is, however, problematic, in that it is more difficult to pin-point which CPU is which... and some of the default line colors are hard to see (although this is user-configurable to a point). I'll try to tweak it and make it more useable.

-S

ORIGINAL: Scott Reams

The results *seem* to show that the Opteron at 2.1GHz is very
slightly faster than the Athlon FX 51 at 2.2GHz at low latencies.

I'm pretty sure this is because you are using lower-latency memory than I am. That Corsair RAM has much more agressive timings than my Kingston memory. Otherwise... an AthlonFX51 -is- an Opteron 148.

Right. But an Opteron 148 runs a 2.2GHz core with a 400MHz FSB,
whereas overclocking the 146 by 5% runs a 2.1GHz core with a 420MHz
FSB. So it has a 5% slower clock but a 5% faster FSB relative to the 148.
Which effect would dominate would be a function of the cache miss rate.