ATI’s Radeon HD 4870 in CrossFire ran circles around NVIDIA’s GeForce GTX 280, but XFX has a GTX 280 in their factory-overclocked XXX flavor. Is the added raw power enough to make the GTX 280 look not so unconvincing against ATI’s dual card solution before NVIDIA’s rumored (not-so-impressive) GT200b appears?

At one point of time, the NVIDIA GT200 was thought to be a formidable chip that noone could touch, but it ended up getting tamed by ATI’s RV770. With ATI’s R700 coming round the corner and the current Radeon HD 4870 512MB making the GeForce GTX 260 896MB look ordinary, there are only two options: overclock the current GT200, or shrink it (GT200b).

Today, we are looking at an overclocked NVIDIA GT200-based GeForce GTX 280 card from XFX, the XFX GeForce GTX 280 1GB XXX. XFX has a tradition of producing cards that are clocked higher than the factory spec, and labelling them XT and XXX versions, with the latter being the fastest, and of course commanding a slight premium.

A quick check at XFX’s web site shows that there are no XT versions for both their GeForce GTX 260 and GTX 280; it’s either you have the Standard, or you have the XXX. Wise move by XFX we would say, having a single highly-overclocked alternative is better than having more overclocked variants as it ends up confusing everyone.


The XFX GeForce GTX 280 1GB XXX comes in a slightly large catchy-looking and contrasty box.

 


‘XXX Edition’ sticker affixed on the box, together with a ‘670MHz CORE’ label indicating what’s inside is a XXX card.

 

This card is surely fast, but will it have enough processing muscle to hold the upcoming ATI R700?

 

Before we continue, below is a table containing the differences between the ATI Radeon HD 4870, the NVIDIA GeForce GTX 280, and XFX’s XXX card.

 

  ATI Radeon HD 4870 NVIDIA GeForce GTX 280 XFX Geforce GTX 280 1GB XXX
Texture Units 40 80 80
ROPs 16 32 32
Stream Processors 800 240 240
Transistor Count 956 million 1.4 billion 1.4 billion
Memory Bus 256-bit 512-bit 512-bit
Frame Buffer 512MB 1GB 1GB
Manufacturing Process TSMC 55nm TSMC 65nm TSMC 65nm
Core Clock 750MHz 602MHz 670MHz
Shader Clock 750MHz 1296MHz 1458MHz
Memory Clock 900MHz GDDR5 1107MHz GDDR3 1250MHz GDDR3

 

Page 1 of 812345...Last »

 

Leave a comment