AMD "Tonga" Silicon Features 384-bit Wide Memory Interface

[Kaapstad]

In what could explain the rather large die-size and transistor-count of AMD's "Tonga" silicon, compared to "Tahiti," it turns out that the silicon features a 384-bit wide GDDR5 memory interface, and not the previously thought of 256-bit wide one. The die is placed on a package with pins for just 256-bit, on the Radeon R9 285, but it can be placed on a bigger package, with more pins, to wire out the full width of the memory bus. This isn't the first time AMD has done something like this. Its "Tahiti LE" chip was essentially a "Tahiti" die placed on a smaller package with pins for just a 256-bit wide bus, on the oddball Radeon HD 7870 XT.

What this means is that AMD's next performance-segment graphics card based on the "Tonga" silicon, could feature 50% more memory bandwidth than the R9 285. The stream processor count is still 2,048, but these are more advanced Graphics CoreNext 1.2 stream processors, compared to first-generation ones on "Tahiti," offering more performance per Watt. The TMU count remains 128, although there's no clarity on the ROP count. Estimates are split between 32 and 48. The R9 285 has 32, and so does "Tahiti."

http://www.techpowerup.com/205811/amd-tonga-silicon-features-384-bit-wide-memory-interface.html

384bit bus and memory compression, this could get interesting.

 

[Kaapstad]

Well, I'm a won't ever buy Nvidia again'er. Ever since I was burned by their defective hardware from their poor choice of solder. Lesson: never trust a hardware company which boasts that they are a software company (seem to recall some JHH quote along those lines).

First to go was my 8800GT back in 2011. Just shortly after BFG went bankrupt. This was followed shortly after by my brother's 8800GT (not BFG). Then came a bunch of other failures of stuff either I or someone I knew owned like 8400M GS, 7150 chipsets and some laptop chipsets. All manufactured around the same time. Nvidia never did come clean about how widespread their problem was, but the wikipedia entry:
http://en.wikipedia.org/wiki/GeForce_8_series#Problems
only mentions G84 and G86 but I don't believe that.

Anyway, won't ever buy Nvidia again. And if all these AMD doom&gloom stories turn out to be true and AMD go bankrupt, I'll just put up with Intel HD. At least Intel stand over their products (even when the problem is minor like SB SATA ports), unlike Nvidia.

As for the thread subject: this would a surprise. Obviously, the chip die size made it look like Tonga was step back in efficiency but if a large 384-bit version with more shaders is coming this might not be true. Would this be first time that AMD have die harvested and binned a chip by bringing out the cut-down version first? I remember thinking since the 7950 GHz Edition that AMD should really do more binning: 7950 GHz Edition @ 1.25V was hot and loud, while a binned version at 1.00V would have been far cooler and quieter. After all, a lot of miners saved 20-40W by doing exactly that.

Go for it.

Yep them NVidia cards are crap, the drivers don't work, they are the wrong colour, the wife hates them, my car broke down, the Titans make fantastic house bricks, look where they got ET, there is a 50% higher chance of slipping on a banana skin and the aliens were using them in Independence Day.

Got any more.