I'm just one of those old-timer enthusiasts who feels like the only guy who remembers things sometimes. I hope others around here can remember one of the most legendary video cards of all time though. This beauty launched 10 years ago on November 8, 2006 (just in time to be a godsend to spend-happy PC enthusiasts for 7th generation console ports and Oblivion):
The G80 GPU powering it was nothing short of revolutionary. It was the first PC GPU to feature unified shaders, was so successful of an architecture (Tesla) and a chip that it was shrunk down into G92-based 8800GT and 9800 series with no true architectural replacement until Fermi 4 years later, and was the first truly big-die GPU (weighing in at 484mm^2). In contrast, the new GP102 GPU in the Titan X is only 471mm^2 (so smaller than G80) and the biggest GPU from any vendor before the G80 was a measly 333mm^2 (G70 - 7800GTX) iirc.
Also, it was something crazy like twice as fast as its predecessor (7800GTX) for pretty much the same price. GPUs advanced pretty darn quick back then and this one was even more exceptional than usual for its unique traits (also black PCB, very rare back then; Nvidia's designs tended to have green PCBs and ATi's had red or blue for Sapphire).
Due to its PCB layout/color and cooler design, 8800GTX was not only arguably the first "modern" GPU as we know it in terms of programmable shaders, but it also looked the part too. Here's a shot under the hood (note: die is covered by IHS like we still have on CPUs; it's not quite that big):
Notice 384-bit bus and memory layout and general placement/look of everything (like power connectors) as on modern flagships. In contrast, here's the 7800GTX (8800GTX's predecessor):
and the ATi X1900 from the same time:
Red, everywhere. The memory layout on the 7800GTX in particular looks kind of funny next to modern GPUs.
So yeah, thanks to anyone who read through this all. I just had to mention this monster's decade anniversary.
The G80 GPU powering it was nothing short of revolutionary. It was the first PC GPU to feature unified shaders, was so successful of an architecture (Tesla) and a chip that it was shrunk down into G92-based 8800GT and 9800 series with no true architectural replacement until Fermi 4 years later, and was the first truly big-die GPU (weighing in at 484mm^2). In contrast, the new GP102 GPU in the Titan X is only 471mm^2 (so smaller than G80) and the biggest GPU from any vendor before the G80 was a measly 333mm^2 (G70 - 7800GTX) iirc.
Also, it was something crazy like twice as fast as its predecessor (7800GTX) for pretty much the same price. GPUs advanced pretty darn quick back then and this one was even more exceptional than usual for its unique traits (also black PCB, very rare back then; Nvidia's designs tended to have green PCBs and ATi's had red or blue for Sapphire).
Due to its PCB layout/color and cooler design, 8800GTX was not only arguably the first "modern" GPU as we know it in terms of programmable shaders, but it also looked the part too. Here's a shot under the hood (note: die is covered by IHS like we still have on CPUs; it's not quite that big):
Notice 384-bit bus and memory layout and general placement/look of everything (like power connectors) as on modern flagships. In contrast, here's the 7800GTX (8800GTX's predecessor):
and the ATi X1900 from the same time:
Red, everywhere. The memory layout on the 7800GTX in particular looks kind of funny next to modern GPUs.
So yeah, thanks to anyone who read through this all. I just had to mention this monster's decade anniversary.
Facebook
Twitter