PIV Vs. XP Analysis

With fairly cheap low-speed Northwoods, it’s pretty much even at the moment.– Ed

Update: Before we start, these low-speed Northwoods aren’t exactly taking the world by storm. I’ve gotten just a few 1.8GHz reports and no 1.6GHz reports.

VR-Zone now has some reports on 1.6GHz processors with aircooling (albeit with RDRAM).

The folks who talked to me about their 1.8GHz chips indicated that 2.4GHz looks like an easy overclock, with little to no voltage increase or super-duper cooling. Getting much past that means pumping up voltage a lot.

The VR-Zone reports seem to follow the same pattern.

It’s still very early in the discovery phase (especially for the 1.6s), but 2.4 seems to be a fairly safe bet for these chips.

Something Screwy Is Going On

As people get past 2.5GHz and head towards 3GHz, some screwy things are going on with the PIV. In more than a couple benchmarks, matters essentially stop improving, indeed, sometimes it gets worse.

Tim Cole spoke about that a few days ago, and a few others have chimed in. Most lately, a few folks at Tom’s Hardware have posited that DDR just doesn’t cut it at high speed.

We’re not so sure that’s the sole explanation, but whatever the reason(s) are, it doesn’t seem like killing yourself to get past the 2400-2600Mhz range does you a whole lot of good fairly often.

Intel or AMD?

Forget voltage mods, forget Peltiers or LN or all the hardcore stuff.

What can the lazy overclocker reasonably expect right now from the two companies?

From what I’ve seen, seems like 1.67GHz or a little better than that is a realistic expectation for an XP.

As mentioned above, 2.4GHz seems a realistic bet with the PIV.

The low-speed Northwoods greatly reduce the AMD cost advantage. You can get a 1.6GHz for about $170, and a 1.8GHz for about $230. This isn’t too much more than XP equivalents.

Performance? While benchmarking up to now doesn’t exactly meet this matchup, it’s probably pretty safe to say in general that Intel holds a small advantage, an advantage that gets even smaller if you can push that XP past 1.67Mhz.

If you have a high-end TBird or XP system, is the advantage enough to change platforms? No.

If you want to get a new system one of these days, is the advantage enough for you to push up your plans and buy now? No.

If you need a new computer now, is the advantage enough for you to go Intel? Ehhhhh. This is a tough call, made tougher by this “why doesn’t this work faster when I bump up the speed” issue.

If these problems are CPU-related, then this isn’t too big a deal. If this is truly a memory bottleneck, then you’re kind of stuck. The future debut of PC1066 RDRAM boards and dual-channel DDR boards sort of indicates the PIV could use more bandwidth.

On the other end, we still don’t know for sure if current mobos will support Thoroughbred. Right now, I can see some reasons why it would, but I can also see some reasons why it might not. If it does, that could be a very strong reason for people to stay with AMD. If it doesn’t, that could be a strong reason for irritated people to go to Intel.

Unfortunately for Intel, I’m afraid the last point illustrates the decisive question: How much do you dislike/are fed up with/etc. with AMD? If your answer is “a lot,” then this is a reasonable option if you need something right away. If that’s not your answer, there’s no real reason to change. If you’re not committed, it’s a coin-flip that may or may not be right depending on current unknowns.

However, it is no longer absurd to pick a low-speed Northwood platform over an Athlon XP platform if you’re starting from scratch, and that represents a change.

Email Ed

Be the first to comment

Leave a Reply