Nvidia Geforce 8600 series

is the 256MB NVIDIA GeForce 8600GT-DDR3 a lower or upper end card for the 8600's? which is the best that is an 8600?

 

There are two 8600M cards, the 8600M GS and 8600M GT. The GT is much better than the GS.

There is a 256MB and 512MB version of the 8600M GT. I've heard stories that each of them are better than the other. I don't think it's much of a different between 256 and 512, though.

 

The 8600 GT cards also vary since they have either GDDR2 or GDDR3 memory. Im guessing the highest one would be the 512MB GDDR3 8600GT.

 

How would you know what had what, like what do the new 1520's have?

 

i was looking at the 1520's and the graphic card spec changed from ddr2 to ddr3 within a couple of hours so id be carefull when ordering and make sure it was the ddr3 version b4 u press the order button

 

i would rank the 8600m gt's this way.

1st place: 512 meg ddr3 gt
very close 2nd place: 256 meg ddr3 gt
3rd place (tie): 128 meg ddr3 gt / 512 meg ddr2 gt

that said- they are pretty close, and all of them would beat out an 8600m gs, some by more than others.

 

Just to clarify one thing-DDR3 memory is not faster than DDR2. DDR3 is built on a smaller process and can reach higher clock speeds. It also uses less power. At same frequencies DDR2 has the same performance as DDR3.

 

so are the 8600m gt 256mb ddr3 models clocked a lot higher than the 256mb ddr2 versions then or are they just the same clock and the ddr3 models just run cooler

 

The DDR3 models are clocked higher.

 

Actually ddr3 has slightly more bandwidth then ddr2.

 

yeah chaz is right, as usual (sort of).

i guess you could imagine the gpu as a little sub-pc. there is the processor speed and the memory speed. processor is constant regardless of ddr2 vs 3, but the memory clocks are whats different here. 700mhz for ddr3, 400mhz for ddr2.

i think that makes it fair game to say "ddr3 is faster," because in practical terms it is. theoretically, you can take control of your own clocks and do whatever you want with them.

but thats sort of like saying a core 2 duo isn't any faster than a pentium 4... just depends on the clocks. sure- you could clock your p4 at 8ghz and outpace a solid core 2 duo, but you will literally fry the components. similarly, you can just clock your ddr2 memory up to 700mhz. again, you will fry it.

 

So basically it should not have much impact on the performance of the card?

 

Yeah, people are just being really picky here. There isn't going to be a substantial different in the performance of all the version of the 8600M GT card.

 

so, from where those the 700 3dmark06 (from 3600 to 2900) came between the asus G1s and a similary configure Dell 1520?

 

Let's talk about FPS rather than 3Dmarks...

 

it might be a difference of 10% or so. maybe a little more.

so 30 vs 33 fps, or 60 vs 66... maybe a little more or less depending on the situation.

 

Ok, so with a high-end card, lets say a 7950 gtx, and an 8600m gt (I know the 8600, is on a low end of that scale, but bare with me), I'm assuming the CPU would be the bottle neck in some games being as these cards can run games quite well (well the 7950 for sure), about what speed proc would you say is a safe speed to avoid becoming a bottle neck? Being as 2.4ghz is the highest santa rose I can find would this proc bottle neck any current games? How about a lower proc such as a 1.8ghz or 2.0ghz?

I'm just wondering how low you could safely go without running into bottleneck situation for some of the current games.

 

just get the fastest proc/gpu that you can reasonably afford.

they both matter. your cpu might matter moreso in future games with heavy physics and ai going on.

right now, any santa rosa processor will do. 1.8ghz is fine. 2.0 or 2.2 is probably the best bang/buck. 2.4 is even better, but overpriced.

8600m gt is technically midrange, best card by a longshot if you want a 15" laptop. 7950gtx is the best card if you want a 17"+ laptop. understand that they are thicker and less portable, but still much cleaner and slicker and much more portable than a desktop.

 

http://en.wikipedia.org/wiki/GeForce_8_Series

it looks there are two types of 8600gts. look at the bottom chart.
the bandwidth can come in either 12.8 MB/s or 22.4 MB/s

it looks like the dell got the short end and got around 12.8 even though their card was advertised to get 22.4.
does anyone know which card the sager2090/comapl ifl90 gets?

 

from what ive been readin though it would look as if the 256mb version is clocked alot higher, and has turbo cache which technically gives it 512

 

Turbo Cache is trash compared to dedicated.

 

yea it is, but if the gddr3 is clocked higher on the 256 version, and the i think the 512 version has ddr2 at 450 or something around there, then the 256 is deff faster, or do i have no idea what im talking about

gddr3 version is 700mhz x 2

 

Most of the time for the applications of today, yes. But who knows two years from now?

 

its rele bothering me why nvidia would do that, ughhhh

 

Nivida doesn't make the cards. They just design the chips.

 

he never said anything about nvidia making the cards

he's just wondering why nvidia would do that/design the cards with different vram

 

well it makes absolutely no sense, i dont see what advantage that is getting them, maybe there the same price, less ram(faster card), or more mem(slower card) possibly better for future games that use more mem?

 

Both are double pumped I do believe.

 

ahh of course, my bad..