I overclocked my car

Well, almost.....I got a specialist to download a custom designed map, to extract around 40% more power from the existing engine, then with a rather fussy chappie sitting beside me; we QA the ECU remapping by taking it on the road and redlining the car in third (and fourth where poss, but this equates to well over 100) to check that fuelling is correct and not going to cause detonation by having not enough fuel.

Doing this, I actually discovered 2 dead sensors, both measuring airflow/content, one in and the other out of the engine which may have killed the car had I been too enthusiastic about racing the BMWs at the lights :)

A good job this fellow is more conscientious than me with my fat heatsink, fcuk off fan and heat transfer compound. After all, a blown chip costs less than a new engine :)

The easy bit was changing the map (5 minutes), the time consuming bit was testing the car (2.5 hours).

It's a bit scary now; it overtakes slower traffic as if they were signposts by the side of the road; and it's far too noisy to keep good relations with the neighbours.

All from an innocuous looking estate car with roof bars, baby seat and no badges :)

Stanley M
Email address supplied

----->

Hi -- I just found your site and have a couple of comments:

- It looks like "the Register"
- The content is like "the Register" was a couple of years ago when it was still good!

Congrats on a great site

Rich Clarke
Email address supplied

----->

Hi ladies and gents. I think it's great that you have a dialogue with these big companies, because it's hard for us little guys to get heard.

I have been a long-time supporter of AMD, having bought as many AMD processors as Intel (which puts me above the average!) I'm currently typing on a dual Athlon box, built by my own hands.

However, I have to agree with Arron's article completely, at least in spirit. It may be true, in fact, that AMD have not changed the standard that they use to measure, but if true, then there is a flaw in it. Because it just seems wrong that a new 3000+ processor should do worse in a significant number of tests than a 2800+ processor, and that is exactly what happened in a number of published reviews.

Maybe the AMD suite of software isn't broad enough, or maybe the criteria isn't quite enough. First, they need to look at some of the software tested where the 3000+ Barton did worse than the 2800+ Tbred, and incorporate that into their benchmark suite, if they're not already using it. Then they need to add to the criteria for determining numbers: If a processor of new design (for example expanded cache) does worse than a lower-rated processor in X% (say 5%) of the benchmark tests, it shall be bumped to at least the same lower number as that processor that did better. For example, in this case, it would have pushed the current Barton 3000+ down to the 2800+ level.

I think AMD over-reached with this one. Everyone was grudgingly accepting their PR numbers because they were pretty accurate. But I think they got greedy. Just a little extra effort, just a little extra cushion, would have put the Barton 3000+ up to the next MHz bump, and secured its position. As it turned out, it's questionable whether it deserves the rating, and that puts in question the whole PR scheme.

Don't get me wrong, I believe the PR thing is totally necessary since Intel chose raw MHz over IPC efficiency -- they are obviously not in sync any longer. But if AMD is to maintain their credibility here, they need to continue being very conservative in these ratings, like they were in the beginning. I see that conservative cushion has pretty much eroded completely, especially when a 2800+ processor beats a 3000+ in a significant minority of the tests independent reviewers throw at it.

As it was, the expanded cache in Barton was a welcome bonus. They should have left it as a bonus, where a Barton 2800+ was everything a Tbred 2800+ was, but with extra cache! But with the 3000+ fiasco AMD pretty much sucked all the advantage right out of it, and in a small number of cases even made it a penalty.

AMD needs to do a little soul-searching and seriously re-think its position on this.

Thanks for listening.

Michael V
Email address supplied

----->

What in the world is Brandon "Toysrme" Noel talking about? I want to know where he heard that the reason people were upset with 3DMark 2003 was because their scores were lower?!

The reason people are upset is because the "new standard for easily and objectively measuring and comparing modern PCs' 3D graphics performance" doesn't even test what it's supposed to. Yes, it does make cards go slower, but apparently Mr. Noel believes that this, in and of itself, is reason enough to lend creedence to its results. Of course this is ridiculous because anyone could write a program that gets low fps, but that doesn't mean that they'll magically be the best predictor of future performane in actual games.

The reason people are so upset with the new 3DMark is because it doesn't do what it is supposed to: give you a general idea of how your current configuration will be at running future games. Your overall score consists of a DirectX 7 game (?!), two DirectX 8.1 games, and one final game that uses some DX 9 features, but not all. How is this an indicator of future performance? The first three games are just poorly written at that. Nobody writes games like this.

The goal of a synthetic benchmark is to mimic a "real" game. 3DMark fails miserably in this respect and thus does not do at all what it sets out to do.

Regards,

Sean Griesheimer
Email address supplied

----->

I don't think they have amputated Barton because it doesn't scale well. It think they did it because Hammer has to be better when it comes. They need to hold Barton back to make Hammer look better.

Bjørn
Email address supplied

----->

I just don't understand all the negativity the graphics card vendors, and review websites are expressing about 3DMark 2003. All sorts of comparisons are being made between 3DMark and 'real world performance'. Maybe I've missed something, but this is BENCHMARKING software, or for those who are confused about what a benchmark is, it's a ruler by which to measure your technical prowess. Who cares if you get a 7 fps pixel shader score on a Raedon 9700 Pro and a 6.8 on a Geforce 5800. The whole idea is that there is some sort of level playing field where everyone uses the same ruler, 3DMark.

I can't believe all the unfounded negativity and suggestions that one vendor had more influence over the product than another vendor. It's a ruler, nothing more.

Tim Dressel
Email address supplied

------>

What I don't get is why UK citizens are getting so worked up about ID cards. Case in point: Hong Kong, being formerly a British Colony, got ID cards shoved down their (our) collective throats sometime during the last 150 years of British rule. As such, all Hong Kong citizens over the age of 11 carry around a photo ID card, and even children aged 8-11 have an ID card.

Point being, nobody in Hong Kong gives a damn about what 'privacy issues' are being invaded. It is a legitimate way or screening out underage drinkers, smokers and arcade junkies (law declares nobody under aged 16 is allowed into a public arcade).

If the British were so enthusiastic about making Hong Kong people carry ID cards, why can't they do so themselves? Bah. Hypocrites.

Andy.

PS: We use our completely-privacy-invaded ID cards as a passport. It's all you need to enter the country. Which brings up the point - if UK Citizens carry drivers licenses and passports, whats the big deal about an ID card?

---->

"Richard Baker,Northern European Marketing, AMD"

Well, Princess, where can I find a mobo for Athlon that supports ECC DDR SDRAM? Hello McFly? The Asus 760MPX based board we *just* used with VERY good results for a client and fully intended on using for our own system upgrade is no longer available! Yet, there is a very tasty Barton based Athlon with 333FSB (soon 400?) that justy begs to run Linux and just begs to be used in the <20 user "non-enterprise but everyone else" segment. VIA? SIS? Bueller? WTF? Hammer may be the cat's ass but there are folks trying to provide (non-Intel) solutions TODAY. Name, email address supplied

----->

I've wondered for a while now if the reason you haven't posted many intelligent letters is because of some of those deep, repressed insecurities. Of course if they are subconscious...that's much more understandable. But then there may never be any hope. I think if you could work on understanding the real you, the benefits could bear some surprising if not pleasant consequences. You could relax more around those whom you would be in disagreement. You know...no need to feel threatened. You would allow, and maybe even enjoy, a challenge to your intellect. It wouldn't all have to come at once. Maybe one or two e-mailers a week who were very polite, honest, and could actually express themselves with some obvious credibility. When the painful insecurities would surface you could more readily confront them and reinvent yourselves. Its not as difficult as I'm sure you think. Our former president Clinton did it all the time. But he became secure with 100% lies, and that's not quite what I'm talking about.

In due time, you could become real people. Smiling in the face of adversity, ready to conquer the terror, and finally living with the reality that you are no longer wimps. Just imagine, if that day ever comes the rest of us can read the letters and know for sure that intelligent people really do come to this site.

Have a bold, gentle, and very brave new day,

Steve
Email address supplied

It is INQUIRER policy not to publish people's email addresses unless specifically requested