UPDATED Nforce 4 tests

jcschild · Posted: Tue Jan 18, 2005 12:48 am Post subject: UPDATED Nforce 4 tests

Hey Guys,
while still in infancy here is some results. this is the 5 towers test
Thonex's coming

this is the same system that bombed with the Tascam 1884.
Asus K8N SLI
1 Gig of ram
AMD 64 939 3000 stock speed
RME MAdi this time
used both N2 2.10 and N3 3.1 both tests were identical. didnt swap new for old magnetos

24ms Stop 43% Play 68%
12ms Stop 45% play 69%
6ms Stop 49% Play 73%
3ms pops clicks
bear in mind this is a AMD 64 3000 939 gonna try it again with a 3500 (as well as try OCing the 3000)
_________________
Scott
ADK

Home of the "Kentucky Fried DAW"

Thonex · Posted: Tue Jan 18, 2005 12:53 am Post subject:

Is this with the "Thonex" Test or Five towers test?... or some other test?

T
_________________
Nuendo 5, Mac Pro 8 Core 2.33 Mhz, OS 10.6, 16 gigs of Ram, MOL , RME 9652, 2xSamsung 21.3" LCDs (1600x1200 DVI), 2 X-Keys.

jcschild · Posted: Tue Jan 18, 2005 12:53 am Post subject:

opps beat me too it.. i thought this was yours...
coming shortly!
_________________
Scott
ADK

Home of the "Kentucky Fried DAW"

Thonex · Posted: Tue Jan 18, 2005 12:56 am Post subject:

jcschild · Posted: Tue Jan 18, 2005 1:50 am Post subject:

well with the Thonex test results same as the Tascam.
even at 23ms no go.
well so much for an update.
_________________
Scott
ADK

Home of the "Kentucky Fried DAW"

Jake68 · Senior Member Joined: 08 Jan 2005 Posts: 2462

I remember seeing a collation of all the results for this test does anyone have a link to it please?
_________________
Intel Quad 2GB Corsair 6400 Asus Commando.
Windows XP Pro SP1a
AMD X2 6000+ Nvidia 2GB Vista Test Box. Low Memory Footprint in Vista64.
2xUAD-1 RME Mface
Nuendo 3.2 1/ Cubase 4.1

composer · Guest

franYo · Member Joined: 06 Jan 2005 Posts: 369 Location: Croatia

efernan · Member Joined: 07 Jan 2005 Posts: 218

Hi,

Jake68 · Senior Member Joined: 08 Jan 2005 Posts: 2462

Alex · Junior Member Joined: 08 Jan 2005 Posts: 50

franYo · Member Joined: 06 Jan 2005 Posts: 369 Location: Croatia

jcschild · Posted: Thu Jan 20, 2005 3:48 pm Post subject:

on a side note i finally got ahold of the GIgbayte Single PCI-E Nforce 4, will be here today!
this is one i am personally gonna do myself. (i neer get to have any fun anymore so its about time)
aside from looking at audio ability its time to play overclocker again.
i plan on doing all the PCI tweaks to see if i can turn off the video hog.
_________________
Scott
ADK

Home of the "Kentucky Fried DAW"

Alex · Junior Member Joined: 08 Jan 2005 Posts: 50

jcschild: I'll eagerly wait to hear how it works out for you and to hear how it benchmarks. I'm really close to pulling the trigger on a new NF4 machine myself.

Franjo: Something seems fishy to me regarding the whole PCI-E sucks cpu power claim (which I've read several times in this forum - not trying to single you out or anything). The benchmarks I've seen on hardware sites of NF3 250 vs. NF4 with the same cpu don't show any major difference. You'd have thought that the FP tests (or any normal benchmarks) would show something if this is chipset problem. Maybe you know something I don't - so please enlighten me! :)

I just noticed that the other higher values reported in the test results on the AMD 64 3000 are for overclocked systems - so my mistake in interpreting the results. The clock speed for an AMD 64 3000 is normally 1.8Ghz if anyone else is trying to figure out which is which.

I'm choosing to wait for the AMD 64 E0 stepping before I buy. There's supposed to be an announcement from AMD later this month about it. E0 reportedly has a better memory controller, a number of bug fixes, far superior SSE2 performance and adds SSE3 support. It also cuts power consumption by around 30% because of a better silicon process. I hope to get a 3200 and overclock it.

franYo · Member Joined: 06 Jan 2005 Posts: 369 Location: Croatia

Alex,

I'm not saying PCI Express sucks CPU power. I am saying (and this is just my guess) that current PCI Express video cards take up huge amount of system bandwidth, clogging up even the wide HT bus (8GB/s in each direction). Not leaving much room for anything else. Btw, according to Scott, it's the same on Intel based systems. Nothing to do with nForce4 chipset.

Franjo
_________________
ASUS A8N-E / AMD Athlon 64 X2 4200+ / SuperTalent 2GB PC3200/ nVIDIA GeForce 6600 / PreSonus FireBOX on PCI-e FW card/ 2xUAD-1 /
Windows XP Home SP2
Cubase 4.0.1.2074

composer · Guest

jcschild · Posted: Thu Jan 20, 2005 11:42 pm Post subject:

a quick note, system is built and loading windows now!

as to the PCI bus hog issue technically what FranYo is saying is correct, however even in the PCI tweak software you can tweak the AGP/PCI-E
cards. i think besides the fact its hoging the HTbus or SMbus on Intel, but the Priority given to the PCI-E is gonna be very high.
leaving little priority for I/O and a PCI hoging Audiocard.

oddly enough at least the Intel board worked. although horrid compared to equally powered P4s with AGP.

heres hoping!

Scott
_________________
Scott
ADK

Home of the "Kentucky Fried DAW"

Tiger Style · Posted: Fri Jan 21, 2005 12:27 am Post subject:

jcschild · Posted: Fri Jan 21, 2005 1:22 am Post subject:

OK,
Gigabyte Nforce 4 PCI-E video
again begining tests, no tweaking of the PCI stuff. basic default bios.
only 512 meg DDR 433 (gonna be trying OCing later)
AMD 64 939 3200

using the Thonex test N2 latest build
RME MADI latest drivers

23ms stopped 72% playing 85% at the 89bar
12ms stopped 90% playing 100% pops clicks

much better than the Asus SLI did, this one has potential! but how much i dont know. to refresh your memory
here is a P4

P4C800-E Deluxe
P4 3.2 (hyperthreading OFF)
2 gig Corsair low latency ram
Lynx 2
XP pro SP1
Nuendo 2.02.35
@ 6msec 85% running, 75% stopped no clicks
@ 12msec not much different no clicks
@ 3 msec couldn't handle it
_________________
Scott
ADK

Home of the "Kentucky Fried DAW"

Alex · Junior Member Joined: 08 Jan 2005 Posts: 50

BrianT · Member Joined: 06 Jan 2005 Posts: 759

If PCI-E video cards are hogging the system, it's most likely a driver issue, not a hardware issue. I'm sure the vidiots developing the drivers know that expectations are very high for these video cards and that gamers (read *framerates*) will be looking very closely at benchmark numbers, which all video card developers cheat on as much as they can without getting caught.

Bottom line is that these video developers will have no problem with choking up the entire system to grab a few extra fps in Doom 3 or whatever, and the PCI-E 16x pipeline is big enough to let them do it like never before. But just because the pipeline is there, it doesn't *have* to be used to hog the whole system. Before long, one of these companies will make a card not aimed at gamers that will lighten up on the resource hogging.

BTW, have any of you calculated the actual bandwidth *needed* to pump lowly 2D video to a DAW? AGP 8x is already supa-mega-gonzo overkill for that anyway. Personally, I don't see much benefit in even an audio friendly PCI-E video card. We're already into a decimal place worth of overkill on AGP 8x.
_________________
Regards,
Brian T

franYo · Member Joined: 06 Jan 2005 Posts: 369 Location: Croatia

ten · Posted: Fri Jan 21, 2005 3:44 pm Post subject:

Check this puppy out.. http://www.nvidia.com/page/quadronvs.html

Nvidia dual head 2d (only?) card, comes in pci-e. I got the nvs 280 version off ebay for £15, as of yet to try it (waiting for the new tyan s2985 to be released) as soon as im all up and running will post you some results.

ten
_________________
- Abit IP35-E - Intel Q6600@3.6ghz
- 4gig ddr2@1200 - 2x 150GB Raptors
- Albatron 8800GT - Dell 3007-HC @ 2560x1600
- Lynx2-C - N4.1/R2.x - XPsp2/vist64 - Adam A7

jcschild · Posted: Fri Jan 21, 2005 4:05 pm Post subject:

Hey Alex,

FYI in the short time i played i got to 2.25 stable OC.
really didnt have time to push it plus OCing these are alot differant than the old days..

i would not run off and buy it just yet. you need to compare what other systems are doing.
the numbers i turned in are horrid! at this time i cant recommend it.
here is a post by David T from the original thonex tests re-post
this is an AMD 939 3500 Nforce 3

192 Samples----5.351msec----63 - 65------69 - 71 %------79 - 84 %------Couple Of Dropouts
256 Samples----6.803msec----60 - 62------65 - 69 %------78 - 83 %-------No Dropouts
384 Samples----9.705msec----59 - 60------61 -63 %------74 - 79 %-------No Dropouts
512 Samples----12.608msec---58 - 60------60 - 63 %------73 - 77 %------No Dropouts
768 Samples----18.41 msec----58 - 59------60 - 63 %------70 - 74 %------No Dropouts

this podunk nforce 4 pales in comparison.

plus the P4 i posted above.

i had to go home last night but will play somemore today.
i am ordering one of those PNY 280's.
and will play with the priorities setting today as well.

good call Brian! and exactly why i am playing with the PCI latency timers, have not really dug into the bios yet and hope i can adjust the
PCI-E bus somewhat.

unfortunately as far as manufacturers are concerned PCI-E hogging is what they want!
for most there is no longer a need for a PCI card in most systems.
even for video editing the PCI-E seems fine. video transfers over the onboard firewire is fine
and even adding a PCI Editing card to the system seems ok. (VT4 toaster which is broadcast level multistream and a bus hog of it own)

Pro Audio is only 3% of the market.

i think single chip systems are going to become harder to do for audio again, requiring alot of painful tweaking.
_________________
Scott
ADK

Home of the "Kentucky Fried DAW"

jcschild · Posted: Fri Jan 21, 2005 8:02 pm Post subject:

ok trying both latency timer in the bios and using Powerstrip has done nothing for performance.

i have little hope for the Nforce 4 boards.
i am going to contact AMD in hopes they may have an answer.
as bad as performance is i dont know that a 2d card will help.
i have one last thing to try which is a bios hack.
_________________
Scott
ADK

Home of the "Kentucky Fried DAW"