/ Sign-up
Your question

Yet another core revision for thoroughbred?!?!

  • CPUs
  • Core
Last response: in CPUs
December 9, 2002 5:42:53 PM

<A HREF="" target="_new"><b>Anton</b> from XBit Labs writes</A>:

<i>According to “AMD Thermal, Mechanical, and Chassis Cooling Design Guide 23794” dated November, 29, 2002, the Sunnyvale-based company plans to unveil Thoroughbred C0 (CPUID=682) sometimes in future. The die size of the novelty is 86.97 square millimetres compared to 84.66 square millimetres of the Thoroughbred B0 and 80.90 square millimetres of the Thoroughbred A0.</i>

It seems as though Revision "C" will significantly narrow the clock speed gap between AMD's thoroughbred and Intel's Northwood considering that Intel will put far less emphasis on clock speed in 2003 as they did in 2002.

<i>It's your world kid!!!</i>

More about : core revision thoroughbred

December 9, 2002 6:31:40 PM

Sounds almost like Barton. I mean increase the die size a little more to add some more cache, and violla, Barton. Either that die <i>is</i> Barton, or AMD is testing out a new core revision to build Barton from is my guess.

<A HREF="" target="_new"><font color=red>Join</font color=red> <font color=blue>the</font color=blue> <font color=green>THGC</font color=green> <font color=orange>LAN</font color=orange> <font color=purple>Party</font color=purple>!</A>
December 9, 2002 8:41:58 PM

It may be a planned migratory rev in preparation for Barton, but I doubt it will be Barton. 2mm^2 is not nearly enough space to fit another 256K of cache on there...

<i>I can love my fellow man...but I'm damned if I'll love yours.</i>
Related resources
December 9, 2002 9:51:13 PM

Possibly some innovations from the old Mustang core never seen? Could be SS2 instructions, data prefetch, improved FPU?

Soon enough, Intel will make the i845s...imagine dual channel Sdram...*shudder*
December 10, 2002 1:37:50 AM

Or it could just be an attempt to get yeilds up.

*Dual PIII-800 @900 i440BX and Tualeron 1.2 @1.74 i815*
December 10, 2002 1:57:19 AM

If this is the C1 of AMDs, this could spell out good news. Some were right, AMD is trying to improve the 0.13m yeild and quality and it seems this might be it. Hopefully that will have better thermals and if possible to really push the clock speed.
A 3GHZ K7 will really surprise me if it's done soon at Retail cooling.

The THGC Photo Album project,<A HREF="" target="_new">CLICK HERE!</A>
December 10, 2002 2:00:56 AM

Just wait till AMD starts rolling out the 1700+'s on the Tbred B core instead of the Tbred A core

Instead of Rdram, why not just merge 4 Sdram channels...
December 10, 2002 7:25:12 AM

Umm.. just looks to me to be a little gate tweaking and path juggling... not unlike what Intel did with the NW a month or so after the 533MHz FSB versions appeared. In Intel's case, it was just a matter of cutting down the die size, and increasing yield per wafer, but it also seemed to have the effect of letting the C1 scale a little better. Perhaps that's the plan, or, as most likely, R&D found out that the previous revisions weren't working well for adding the extra 256Kb L2, so a little transistor shuffle needed to be done. Ultimately, AMD would like to be able to just slap on the L2 transistors without having to do a major core remap.


Some day, THG-willing, I shall obtain the coveted "Old Hand" title.
December 10, 2002 11:10:13 PM

<A HREF="" target="_new">;/A>

Advanced Micro Devices (AMD) recently decided to introduce its 0.13-micron Thoroughbred B0 core to its entry-level Athlon XP 1700+ and 1800+ processors to cut costs further as well as improve product performance. The Thoroughbred core has two versions. The slightly updated B0 version features a relatively larger core size and was originally used in a few of AMD’s top-end products. With the introduction, AMD’s Athlon XP processors will now all be based on the same Thoroughbred core.

This might be old news, but it looks like Eden was right, the extra core size was indeed for better yields and improved scaling. Good deal, moving the Tbred B cores to the lower end 1700+ and 1800+. Nice call Eden

Instead of Rdram, why not just merge 4 Sdram channels...
December 11, 2002 12:25:47 AM

By the way... ive seen conflicting reports on the possible Barton shape.

Initially the shape was supposed to be square, take the die length of tbred and make a square.

Just recently i saw a pic of an extended tbred, supposedly a barton, and i dont think it was photoshopped. this cpu was very rectangular a tbred and added a bit on the end.

So whats it to be? Square or rectangular?

<A HREF="" target="_new">;/A>
<A HREF="" target="_new">;/A>
<A HREF="" target="_new">;/A>
<A HREF="" target="_new">;/A>
December 11, 2002 12:56:49 AM

I don't think the Tbred is a square though, i'm pretty sure I saw a picture of it being rectangular, like all the other Athlons. Well, at least according to THG, all the Athlons since the Palomino have had a width (or length if you look at it differently) of 110mm, and varying lengths, with the minor exception of the new Tbred of some 86mm (TbredC?) The Barton is roughly a square from guesses, if you take the transistor difference from the Palomino going to the Tbred A, and then the TbredA to the Barton. About 110mm by 110mm, might be a little more or less though.

Instead of Rdram, why not just merge 4 Sdram channels...
December 11, 2002 3:15:51 AM

Thats the thing.

Toms estimation of barton would be square.
But ive seen a pic of an apparently mobile barton demonstrator chip thats definately NOT square.
Like the tbred, rectangular, but even longer by about 25%

<A HREF="" target="_new">;/A>
<A HREF="" target="_new">;/A>
<A HREF="" target="_new">;/A>
<A HREF="" target="_new">;/A>