Cutting Edge

Cutting Edge wonders whether G4 clock speeds and hot computers are connected, and speculates on a possible way forward for the Mac OS.

I was recently called in to give advice to a television studio about which editing system to buy for high-definition video. HD video is roughly analogous to high sample-rate 24-bit audio, but the data rates are astonishing, at nearly a thousand times the rate of CD-quality audio; that's around 140Mb of data per second. You need eight SCSI drives in a RAID level-zero 'stripe' to achieve this type of performance, and even then you can see from the drive activity lights that the things are working very hard indeed.

So you'd probably expect that any technical problems with an editing system would be caused by the high data rates and the fact that the host computer has to run flat-out just to put a picture on the screen. But no. The problem, it seems, is that the computer gets too hot.

The video editing device I was considering for the studio was the Pinnacle Cinewave. And the computer? A Mac.

Cinewave is the first video-capture device that works with Final Cut Pro on a Mac at high-definition TV resolutions. I remember that there were surprisingly few technical issues when the device first came out, and I'd heard mostly good things about it. There isn't much demand in the UK for HD TV because it isn't a broadcast standard. In the US it is, and there is much more of a buzz around it. It will come here eventually (probably as a premium service on a satellite channel) and when it does I'll be very pleased, because it looks fantastic. Also, the improved production values might possibly provide more work for composers, because superb video demands superb soundtracks. (Well, we can hope!)

I suggested a Cinewave system to the studio because it seemed an obvious choice, given that the alternative they were looking at was 20 times the price, albeit somewhat more powerful. They decided to go ahead with my suggestion, so I started to do some background reading on the Cinewave user groups. The good news was that the product seemed well sorted. The bad news was that it didn't appear to run on the latest Macs!

Is the inside of a fast Mac too warm for comfort where Pinnacle's Cinewave is concerned?

So I wasn't surprised to find that there were issues with the latest generation of Macs. But I hadn't expected to find that the oneissue was overheating.

The Cinewave board has Pinnacle's Hub 3 'memory-centric' video processor on board. It's a custom-build chip that generates a lot of heat — just like any chip that works hard — so it doesn't help if the computer it's in can barely keep itself from meltdown. And it seems that the latest-generation Macs are just too hot for the Pinnacle device to tolerate. (I have now heard that the Cinewave is working with the newest Macs, but that Pinnacle are still advising extra ventilation.)

Now I don't think you could accuse Apple of being careless with their design. It's very unlikely that they simply got the ventilation wrong. No. I think it's symptomatic of a deeper problem: perhaps their famous 'Supercomputer-on-a-chip' processors just won't go fast enough. The problem seems so severe that one explanation, in particular, is looking at least plausible: that they are having to resort to 'overclocking' their processors (making them go faster by simply turning up the clock speed), and living with the consequences, the most obvious of which is more heat.

By the time you read this, you should be able to buy 3GHz Intel Pentium 4 processors. Pentiums running at 4.7GHz have been demonstrated in the labs, and you can easily overclock current Pentiums to 3.3GHz. AMD are not far behind with their cheaper range of Athlon chips. What's the fastest Mac processor? 1.25GHz at the time of writing.

Clock speed isn't everything. But it certainly is something, and the more Apple fall behind, the more frantically they try to convince us on their website that their slower chips are really faster. Yes, OS X is a software miracle. It squeezes every ounce of performance from the G4 chip with its clever scheduling and multitasking. With a 64-bit PCI bus and a sustained throughput of 266Mb per second, together with a workstation-like architecture, there's no doubt that Apple have built a phenomenally efficient machine. But now it's time for a better engine. For every claim on the Apple website that applications like Photoshop run faster on a Mac, I've seen benchmarks run on fast PCs that appear to contradict these claims. I haven't done the tests myself, so I'm not in a position to confirm this, but I have used some of the latest 'twin Xeon' PCs and they certainly 'feel' faster than any Mac I have used. (A Xeon is a type of Pentium processor that is designed to be used in multiples. A twin Xeon PC is analogous to a Dual G4 Mac.)

So where can Apple go from here? They've got great design, a fantastic operating system and customer loyalty that most religions would envy. But they need a new processor. What they certainly don't need is me and hundreds of other journalists telling them this. But in the absence of any kind of road-map, and absolutely nothing else concrete to report on, what else can we do but speculate?

The funny thing is, Apple have been here before. Several times. They've shown that they can successfully port their operating system to new platforms, and persuade their developers to come with them. And it's always been worthwhile: the benefits of platform migration have always been apparent. A year ago, at the beginning of 2002, there were reports in the computer press, and web sites such as www.osopinion.com, of a new 'G5' chip that boasted an even greater number of 'supercomputer'-type features. Interestingly it was slated to run at between 800MHz and 2GHz. You can make the comparisons with Intel's offerings yourself. Today, you'd think we'd have more concrete details, if not products sporting the chip itself. But we don't. Of course we may have tomorrow: you can never tell with Apple.

The trouble is that now is really not a good time to have another "we're moving to another platform. Again." conversation with the makers of Mac OS applications. Commercially and financially it's never a good time, but this time it's really bad, because most developers have been flat out for the last couple of years converting their software to run natively under OS X — when they could have been adding new features or tweaking their software in other ways.

For the last year or so I've been hearing rumours about Apple developing a version of OS X to run on Intel chips. Certainly, the 'Darwin' core of OS X exists in an Intel version, and Apple even documents it for developers (http://developer.apple.com/darwin/news/2000-04-05.html).

Windows is not the only PC graphic OS, as the Linux-based Lindows sets out to prove.

Imagine what would be involved if, say, Direct X was ported to another platform. It would have to be re-written from scratch — the kind of issue that acts as a reality check for anyone suggesting that Apple could easily move to an Intel or any other platform. To do so would be painful to developers, confusing and frustrating to loyal customers, and possibly one call too far for Apple.

So, either way, these are difficult times for Steve Jobs' baby. If Apple don't come up with some faster machines soon, they will lose customers to Windows/Intel. Sales of iMacs are already below targets. If they move to another platform, developers may desert — the ones able to survive at all, that is. What a waste for such a great product.

But maybe there is another way: a third way, if you like.

What I've just suggested might sound like a very messy solution, but it's not as daft as it sounds. As a Windows user, I love the look and feel of Aqua and the Mac OS. I'd love to be able to use it, but I can't re-invest all the money I've spent on Windows programs over the years. It would be great if I could wrap a Mac-type interface around them. This doesn't get around any of the difficulties I've written about above — and I can't begin to imagine what the legal implications would be — but I think it would be a great halfway house to being able to run Mac programs natively on PCs. It would, at the very least, give Apple application programmers a breathing space (although I guess they'd worry about losing business to Windows application programs until their new Apple/Intel versions were ready).

If all of this sounds daft to you, I readily admit that it does to me too. Or, at least, it would if it weren't for Lindows.

No, it's not a typo: that really is Windows spelled with an 'L'. Lindows is actually a version of Linux, incorporating some clever stuff that allows some (but only some) Windows applications to run under it — without a single byte of Microsoft operating system on the computer. ZDNet news reports that Microsoft Office runs happily on it. However, I can guarantee that Cubase SX won't!

Still, it does give some idea of what might be possible, doesn't it?

Home | Search | News | Current Issue | Digital Editions | Articles | Forum | Subscribe | Shop | Readers Ads

Advertise | Information | Links | Privacy Policy | Support