Architosh Staff (info@architosh.com)

1 Feb 2002

 

Pro Mac Survey: Apple's Users Demand More

Architosh has gathered substantial data on the hardware requests of professional Macintosh 3D and visualization users working in a number of industries from film and special effects to medical 3D.

In a related report, as part of a series, users are in hot demand for Quad Power Macs and even an Apple solution for Plug-Play compute/renderfarm distributed computing technologies. Such users say they are willing to pay for them if Apple can deliver the goods. However, these same users chime in on Apple's shortfalls in this market segment and suggest solutions.

More Power Please - A Lot More!

While Apple has clearly ratcheted up performance with a dual 1-Ghz G4 Power Mac, users say they want four of those processors in one machine, not two. But in addition to the number of processors, users acknowledge that Apple needs to do a lot more to entice users away from their troublesome Windows NT machines and expensive and limited SGI and Sun boxes. They see Apple's new UNIX-based OS X as the key to their future...if only Apple could deliver the power.

Andrew Yang, a development specialist in the Advanced Products Group of Stuart Energy Systems Inc, wrote that not being able to deploy the Mac in their environment came down to several factors: 1, lack of high-end solid modeling software such as SolidWorks, Unigraphics, Catia and Pro/E (most of those are UNIX-based); 2, lack of workstation-class graphics cards like those suggested in our survey; and 3, processor speed in conjunction with bus speed.

"It's very apparent that 3D apps benefit greatly from high bus speeds and processors speeds, and the Mac has neither. Not speaking about the CPU, the bus speed has been sitting at 133MHz while Intel is likely to be at 533MHz before the middle of the year." -- Andrew Yang, Stuart Energy Systems Inc.

While such comments are discouraging to hear from UNIX and PC users wishing to make a jump to a robust OS X operating system, Apple also faces a value proposition problem for many hard-core 3D users looking for pure bang for the buck in 3D performance.

"I have to mention that I've made the same scores on visualization when benchmarking with Cinebench 2000 on the following systems :

- Macintosh G4 450 bi-processor with Radeon Mac Edition (32 MB DDR) (more than $3000)

- Home made PC PIII 733 + geForce II MX 200 32 MB (around $ 950)

So we've trashed all our Macs and bought $2000 PC's for all 3D related stuff..." -- User name withheld

In a similar vain, such large studios such as Dreamworks have said they are moving to Linux on cheap Intel hardware in lieu of expensive SGI, Sun, HP or IBM UNIX boxes. Clearly UNIX/Linux is where the future lies and more studios will likely move over to such pipelines. The question becomes where does Apple fit in this value-chain reaction?

Many survey respondents mention that key design and video software like Photoshop and Final Cut Pro not only "not run on UNIX" but run best on the Mac. Would Adobe ever ship Photoshop for Linux? And how would such moves affect both Apple and Microsoft?

Clearly Apple is in a supreme position over Microsoft in this value-chain reaction. Because Apple's Mac OS X is UNIX and Windows is completely not UNIX, the portability of UNIX and Linux software to OS X is extremely promising. However, Apple's current expensive hardware is in a currious middle point. Much less expensive than SGI and SUN gear, yet more costly than Dell PC's.

By now it should be obvious what Apple should do and what it has to do.

Apple Can Do It

A number of readers said they have great faith in Apple (faith in Apple...imagine that?) and believe it is only a matter of time before they clearly deliver the goods. According to some astute readers, they have the technology now to do incredible things.

"While serious graphics cards are a crucial _entry-point_ barrier for OSX, the real issue at the other end of the spectrum is scalibility." -- David Faibish

David Faibish talked about cluster and compute/renderfarm efforts on PowerPC being centered on Linux not Darwin (where they should be). Such efforts are being made by Project Appleseed and Beowulf.

Apple needs to use Darwin to show off MACH to demonstrate scalibility (the bsd is just what drags in the linux codebase); all of the Linux clustering (eg Beowulf etc) use a really lame clustering technology ("MPI" - a brittle, awkward, procedural, IPC glue to kludge together cpu's - which is typically transported via TCP on Gig Ethernet); Apple needs to use multiple optical FireWire channels (each at 3.6 Gbps) and MACH's object-oriented messegebus as a different _programming_ architecture to run circles around Linux/MPI clustering ... -- David Faibish

Apple's rumored "Gigawire" technology is said to possibly be a multi-channel, accelerated FireWire technology (possibly wireless). And their Advanced Computing Group is known to be working actively on parallel computing technologies. David says, "When Apple's MP'clustering technology comes built-in, the whole compute/renderfarm business segment will just explode (and put a big dent in SGI's viability)."

While we see some of what David mentions happening, we see SGI continuing to be pushed up-market (the eventual Northeast migration on the chart), out of reach and threat by Apple.

Apple is Doing It Now

While some users see Apple not being able to compete, and others seeing Apple not taking advantage of technologies they clearly have now, some say Apple's problems are a myth or perception issue. A petition respondent named Johnny had some comments about real life tests:

"It's a myth that the Mac graphics aren't up to pro standards. Do some serious testing of GeForce3 in a G4 running Maya or Lightwave[7] compared to a GloriaDCC running in a Windows2000 box and you'll discover that it's a MYTH that the Mac needs some stupid expensive graphics card. Just run some tests, you'll see!!!" -- Johnny

Closing Thoughts

While these are just some of the comments made in our surveys for more 3D power on the Mac platform, there are many others just like them. Some common thoughts are that Apple "can" deliver the goods but chooses to focus too much on product design (cases) and not enough on its software and hardware design expertise. In particular, the mention of OS X's core strengths are items Apple is not fully utilizing -- such as optimized support for four or more processors (better threading).

Readers often cite amazement at Apple's motherboard progress, or lack thereof. And they suggest that if Apple is lacking the ability internally then they better go out and buy that ability. Some suggest SGI and their high-bandwidth and graphics expertise as natural candidates for Apple's $4 billion. Others suggest dumping Motorola and embracing AMD or Intel. Many wonder why Apple doesn't just buy Nvidia or 3D Labs or some other player. And where is the fruit of the Raycer Graphics buyout?

The common thought is that Mac users have waited since the early 90's to have the next-generation Mac operating system of the future, and now that it is here Apple's hardware is letting it down...and everyone with it. In the past, Apple's hardware made us all very proud (smoking Pentiums and so on), while we knew full well the Mac OS underneith was getting very creaky.

Now we have our new OS. So users are asking, "where's the beef?"

Related Article

Pro Mac Survey: 3D Mac users overwhelmingly want Quad Macs