vacancies advertise contact news tip The Vault
facebook rss twitter

Review: AMD Athlon XP2600

by Ryszard Sommefeldt on 21 August 2002, 00:00


Quick Link:

Add to My Vault: x


Cast your mind back to the 10th of June, some 2 months ago, and you'll remember that AMD released the latest revision to it's Athlon XP line of processors. With a die shrink, voltage drop and consequent drop in heat and increase in operating speed to 1800MHz, XP2200+ represented a major step for AMD. It was their success in bringing a 0.13u part to market after what seemed like an age.

While it represented a big step in terms of their move to a new die process size, we commented that although the new CPU was fast, it could do with a jump in front side bus speed and memory bandwidth to keep up with the new Pentium 4 processors that were appearing.

So AMD were able to execute at 0.13u without much issue, despite the CPU taking its time reaching retail channels (it hit the big OEM's first, at least here in the UK). We were able to push our XP2200+ sample to around 2GHz using air cooling but it seemed to stop there. The freon brigade don't do that much better with XP2200+ with the odd 2.1GHz+ rig making an appearance but for the most part, AMD's new baby stopped at around 2GHz.

This was slightly worrying since the same process switch for Intel with the Pentium 4 Northwood processor led to massive percentage overclocks and a general willingness for the CPU's to go high from day one. Intel have followed through with this and the 0.13u P3's have scaled excellently and we should see a 2.8GHz example this week from them.

But with AMD's new baby hitting a brick wall around 2Ghz, it didn't bode well for future Athlon XP CPU's using the same process. We've heard speculation that we wouldn't see another desktop CPU from AMD until Hammer and that they were changing the front side bus speed but keeping the MHz ceiling below the observed 2GHz limit since that was all they could feasibly do.

Well AMD would love me (and everyone else covering this event) to let you know today that neither of the above observations are true. Today marks the release of a pair of processors, XP2400+ and XP2600+ running at speeds of 2GHz and 2.13GHz.

No voltage increase needed (still the same 1.65V operating voltage for the new CPU's, the same as XP2200+), running cooler than before and hitting much much more than 2GHz (even on air), these new CPU's have arrived in a big way.

While still tied to a 133MHz processor bus and the 2.1GB/sec of memory bendwidth that means, these new CPU's are a full 200MHz+ clear of XP2200+ and a good jump or two in the PR rating scheme that you'll definitely be familiar with when talking about Athlon XP's.

The old formula that meant starting at 1333Mhz for XP1500 and for every 66MHz added to the clock speed, add 100 to the PR rating is broken. Using that formula, XP2400+ should have been 1933MHz and XP2600+ should be another 133MHz down the road at 2066MHz. Somewhere along the line, 66MHz has been added to the initial frequency base of 1333MHz so the new ratings aren't exactly to the same scale as before, but their meaning certainly hasn't changed.

You'd still need the equivalent of a 2.6GHz Thunderbird Athlon to match the performance of this 2.13GHz Athlon XP2600+ according to AMD. So a slight adjustment in the PR rating scale for these new CPU's.

So how have they overcome the previously observed ~2Ghz air cooled ceiling?