• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Nforce Benchmarks you may or may not have seen

Yep - I accidently posted this in CPU forum too (didn't see this post until I went through the trouble of doing the following):

http://www.hkpcug.org/articles/computex2001/day1/day1-1.phtml

I can't read it 🙂( ), so I don't know the whole system configuration - but still, the chipset looks awful promising!

I'll copy/paste the info here:


Stream D Copy64
VIA-KM133| AMD760 |nFORCE-220 |nFORCE-420D
259 557 ....681 970

Stream D Copy64
VIA-KM133 AMD760 nFORCE 220 nFORCE 420D
......259 ............ 557 .....681 .........970

................................................Sysmark |Winstone (Business) |Winstone (Content Creation)
nFORCE 420D + Athlon 1.2GHz .....238 ........45.5................................ 53.7
i850 + 1.5GHz Pentium 4 .............200 ........38.1................................. 47.6



3DMark 2000
...........................nFORCE-420D| nFORCE-220 |i815 |KM133
1024x768x16bit .................4252 .........3378 .......792 ....708
1024x768x32bit .................2683 .........1813 .......0 ......278

Quake III (1024x768x16bit)
AMD-760+M64| AMD760+MX200 |i815+MX200 |nFORCE 220 |nFORCE 420D
.....30.3 ..............51.5 .......................51.5 ........52.9 .............66.8
 
Am i right in assuming the Nforce 220 is the MX 200 Geforce chipset.And the Nforce 420d is the MX-400?or is it GTS?Hell im just impressed in seeing an integrated board that can game pretty well!
Nforces for the whole family this holiday season!
edit: Provided the price is right😉
 
The 220 is the 64bit, single DDR channel. This means that the graphics chipset shares the bandwidth with the processor bus (2.1Gb/sec shared between graphics and processor).

The 420 is 128bit, dual channel, however, the processor still only has a 2.1Gb/sec bus (266Mega transfers/ second, 64bits wide), so the dual channel bandwidth is used primarily for the grahpics. Though I don't know the specifics as to how it works, I suspect that the dual channel will help lower latencies to the processor to some degree (I've not read the tech docs, so I don't know), but the extra bandwidth afforded by the chipset will only be useful in gaming situations (afaik).
 
From what i hear, no one can reproduce the benches of the Nforce though. Maybe that was incorrect, i hope it is a good board, since I want to get a laptop in the future and the Nforce is supposedly gonna move to the mobile market.
 
No their both based on the same graphics core. However the 420 has a Dual 128bit DDR SDRAM interface, while the 220 has a sinlge 128 bit interface or was it a dual 64bit interface (I don't recall).

Thorin
 
Here is a bit about the nForce I posted Here

Actually the 128 bit path is helpful even though the cpu bus is only 64 bits. Sequential writes can be paired 2x64, doubling the memory write bandwidth. Although only 64 bits at a time are sent to the bus, it will take only half as many writes at 128 bits, giving greater performance via efficency. Also, sequential read bandwidth will be increased in the same manner - the controller pulls down 2x64, and if the next 64 bit chunk in on the controller (which will happen in a sequential read), then it can be sent to the cpu immediatley from the controller instead of being requested and read from memory. Although neither reading/writing 128 bits in this manner will increase the peak theoretical bandwidth, it will significantly increase the actual memory bandwidth in both situations by having the next 'chunk' hot and ready to go. This efficency also allows other devices to access the memory more quickly as well as benefit from the same advantages in a sequential read/write mode, such as memory controllers/gfx cards/etc.
 
😉 I must say I appreciate it when senior members show a little respect. It happens far too infrequently 🙂

What many people in the DDR/RMBS bandwidth war fail to realize is that peak bandwidth numbers only give you a misleading sense of performance. What is important is the actual or effective bandwidth when the system is under a load - this is when a more efficient system can take the crown from a system with more peak bandwidth.
 


<< What many people in the DDR/RMBS bandwidth war fail to realize is that peak bandwidth numbers only give you a misleading sense of performance >>

Agreed. Though the Itanium has only 2.1Gb/sec theoretical bandwidth, it has a rather impressive ~1.4Gb/sec STREAM bandwidth. That's ~67%! Most chipsets in the x86 arena are doing rather somewhere wouth of the 67% figure, usually even less than 50%!
 
Looks OK if you're gonna use the integrated video, but keep in mind that a 3D Mark 2000 score of 4200 is pretty lame, even if it kills other integrated boards. What interests me is the Sysmark and Winstone scores, and as you can see, it's nothing exceptional compared to AMD760. Looks like SiS will be the one getting my money.
 


<< wow 4200 on 3dmark, that kicks my gf2mx's bootwah >>



Those are 3D Mark 2000 scores. I'm pretty sure an MX can do 4200, I got over 7000 with a GTS...
 
I am thinking someone made a mistake and those are 3dmark2001 numbers...My radeon ddr in 3dmark2000 on a 750mhz tbird kt133 board would get 5400 and I am sure the gf2 mx should be close to that. Hell look at reviews for the card it does quite well in 16bit mode...32bit and higher resolution is where the gf2 is supposed to lag behind the radeon...that 32bit number looks just like what I have seen some pppl post with those cards, maybe a bit lower though.
 
These benchmarks look pretty sexy but who knows if they're trustworthy since they were done by nVidia. Not exactly an independent third party! 😉

On Tom's site, the athlon 1.2gig is ~8% slower than a P4 1.5gig on the sysmark 2001. According to the nVidia benches however, the nforce 420 + athlon 1.2 is 19% faster than the P4 1.5!!! This is phenomenal!!

But is this for real? Tom's benchmarks say that the P4 1.5 only scores 149 on the Sysmark whereas the nVidia benchs claim that the P4 yields a sysmark of 200. What is behind this discrepency? Who knows.
 
I tend to be incredulous myself. However, Nvidia has delievered in the past with their graphics products. Lets hope they keep on delivering!
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top