Incredible nForce Benches

[MiKiMoNo]

Hello everybody!!

I've been searching all the web behind some benchmaks for nForce, and I've found one almos incredible. It says that you can copy a file of 1.44Gb from one HD to another in 70 seconds (about 20 MB/s) with nForce sample, while AMD760 do the same task in 210 seconds: Three times more!!!

Can someone explain it???

Here is the link to the benchmark: http://www.inqst.com/articles/nforce/0605main.htm

What I'm saying is: How can copy a file in 70 seconds and other chipset (760) do the same in about 210 seconds???

Can we say that current chipset CAN'T hold 20MB/s!?!?!?!?! wait, wait, wait... it's imposible!! or not??

The maths says: If you can transfer a file three times faster with nForce, and its transfer speed is 20 MB/s, then the original transfer speed of a AMD760 board is... ¿less than 7MB/s?
Oh my God!! no no no... I don't believe it!!!

Please, can some one explain that????

thanxs in advance!!



Here is the link to the benchmark: http://www.inqst.com/articles/nforce/0605main.htm

 

[downhiller80]

Possibilities:

1) lies lies lies. All lies.

2) stuff-up in the lab. Forgot to install 4-in-1s or something. Didn't enable DMA etc. Many possibilities when a whole MB has been replaced.


- seb

 

[mk52]

I d trust OJ more then InQuest

-MeliK

 

[downhiller80]

6.4 GB/s on the X-box is 200MHz DDR!!

- seb

 

[Wingznut]

I was thinking something very similar, Melik.

Nice sig, btw.

 

[downhiller80]

I'd like to make a suggestion here. Two sockets on the motherboard. Why? One for the graphics chip....

- seb

 

[bjc112]

<< I d trust OJ more then InQuest >>



LOL,

i still think that would be a nice comb, good performance though!!

 

[zephyrprime]

I've always found hd performance in real life file copy throughput to be nowhere near benchmark throughput. I always thought this was due to crummy windows performance. Could it actually be due to the chipset??

Also, if this was a file copy test, then the actual throughput is 40meg/sec because each meg that is read from one drive has to be written to the other.

 

[zephyrprime]

I jest went to the inquest site and it says that they bothched the previous benchmark graph. nForce only 26.6% faster at file copying.

 

[Maverick]

yeah that can't be right. Disk access time is still the biggest bottleneck in systems. Even the fastest SCSI drives running at 10,000 RPM can't keep up with the system bus. If you're transferring to and from memory, the story changes.

 

[HaVoC]

Inquest has an extreme anti-Intel bias giving all of their benchmarks about the same credibility as Dennis Rodman preaching the Gospel.

 

[nortexoid]

aye, just checked it...only ~26%...

still though, that's a considerable amount...if true, i'm creaming..

will this be my next upgrade?...i hope the 2d isn't god awful.

 

[Spotch]

Well I would applaud but I just did it in 96 seconds on my CUSL2 PIII 700E @ 933
2 x Maxtor 30GB 7200. (While brousing forums and playing an MP3 in Media Player 7 with graphics.)

 

[Eug]

Big load of bull.

 

[Spotch]

78 secs without anything running... My Win 98 install is over a year old too! (Bloated)

 

[p0tempkin]

UPDATE 6/10/01: We have corrected the chart above that previously showed a much larger performance delta between these two platforms. The performance difference here is attributable entirely to the PCI bus that separates the south bridge IDE controllers from the north bridge (CPU and DRAM) on the 760 platform. Today, maximum disk buffer transfer rates are about 100MB/s. This exceeds the typical sustained bandwidth limit of PCI (about 75MB/s), but does not exceed HyperTransport, VIA?s Vlink or Intel?s Hub-Link. We also expect to see fast HD file copy speeds from Vlink and Hub-link platforms.

 

[BlackWob]

I thought nForce only exsisted on paper right now? Like as in there wouldn't be test boards out for a few months? Is nVidia ahead of schedule or is some site just trying to get visitors?

 

[MGMorden]

There are definately some actual chips floating around. All lot of the sites I've been to have shown test boards from MSI, Gigabyte, Asus, and Abit already.

 

[bluemax]

Maybe I should be speaking to board manufacturers to get a test unit.... I've been speaking with nVidia who won't release an evaluation unit until the &quot;testing period&quot; is over. Once it becomes a full retail part, I should get one.

Hope that's soon.... *panting*

 

[thorin]

Well I'm note sure about InQuest and/or their article however from that and other stuff I've read the nForce is very impressive. It is the first &quot;integrated&quot; solution I have ever considered actually purchasing. The i8xx integrated series as well as Via and SiS integrated solutions have never interested me and always seemed to underperform.

Thorin

 

[zephyrprime]

I just thought of something. Inquest says that the nforce is ~26% faster at file copying vs. some other motherboard. But www.storagereview.com shows that different modern motherboards and IDE controllers already in the market have very different IDE performance. So nforce being 26% faster than some other motherboard wouldn't be impressive if the board tested against were on the lower rung of motherboard IDE performance.