HOME » | FORUM | CONTACT |

.: Technology Bytes XVIII:
Prepare for Battle: Attack of the Techno Babble Bunch?!

July 26 2003:

As the dust settles from last months hysteria, an uneasy quiet has descended as the 3 Tribes take a breather and prepare for the oncoming battle which no doubt will erupt over the next few months.

But there's still a bit on....

Chipzilla gained some extra mileage in the Workstation / Server Market by releasing the New 3.06Ghz Xeon DP with a 1 Mb L3 Cache , much to the dismay of the AMD community, and also found time to put a few noses out of joint with reports that they maybe "up to their old tricks again" with the Prescott possibly not being compatible with current i865/875 Motherboards as previously believed. AMD remained consistent at least by reporting yet another loss for the past quarter, but overall kept their heads low, and Apple continued on their blind monkey marketing approach, ignoring the fallout from last months G5 Benchmark debacle to arrogantly maintain the line of having the first (64 Bit ) and fastest box on the block. Sigh :-)

Lets dig in !

Intel raised the stakes in the Workstation and Server Markets on July 14 by releasing the 3.06 Ghz Xeon DP processor that I had reported on a few months back.
This also coincided with a price drop of about a third to the existing Xeon DP range, which is no doubt aimed at stemming any possible flow to the Lateron solutions offered by AMD. The units are built on the New Gallatin Core, but still run on a 533 FSB. ( Units with 800 FSB are not schedule until next year with the 90 nanometer Nocona chip, which is the Workstation Version of the Prescott.) A 3.20 Ghz version is also slated for later this year.

Recent benchmarks on a host of 32 Bit app's have the 3.06 Ghz units easily out performing the 64/32bit AMD units , and lets face it, with the vast majority of real world applications still running 32 bit, and will be for quite a while yet, Intel may just have stolen the AMD64 thunder even before it gets a chance to get rolling.
Here

There are still even more damning benchmarks with the Single Processor Opteron Solutions being slapped around by the PIV -C and the Athlon XP. More on that a little later.

Intel have also pushed forward the release of the 90 nanometer version of the Pentium M chip code named Dothan to late Q3 or early Q4. The chip is expected to ship at 1.8Ghz, will have 2MB of L2 cache and use a 400MHz system bus. Now if we could only get Intel to release some viable chipset solutions to allow these units to also be used on the desktop, we could really be talkin turkey. There has been some initial development for the OEM Blade server market using the existing i7505 chipset, but no real attempt at offering the chips to the mainstream desktop or workstation market.

It hasn't been all smooth sailing for Chipzilla though, with reports here that the current crop of i865/875 motherboards may in fact not be compatible with the coming Prescott CPU's despite the assurances from both Intel and the Motherboard manufacturers. Seems Intel maybe up to their old tricks, and have decided to play "Change the Socket" ( as apposed to hide the sausage..) bringing back memories of the debacle with first the Tualatin 370 CPU's , which had a slight pin layout change, and also the shift from 423 to 478 Pin sockets on the PIV. In both cases end users were forced to also upgrade their motherboards when wanting to upgrade their processors, which left a lot of end users less than impressed with the wisdom of the Chipzilla brain trust.

It is still not entirely clear on what the actual situation is due to the Brain Trust not " officially " commenting on unreleased Product, so we have no choice but to wait it out, and hope that Chipzilla don't leave the current crop of i865/875 Motherboards with a shelf life of less than 6 Months. Mind you , the 478 Prescott CPU was always going to be a transitional phase any way, with the 775 Pin layout being adopted at the end of Q1 2004 with the Grantsdale chipset , but there was an understanding that the current crop of Motherboards would at least see us thru to then.

Ah, looks like we may have been stitched again !!

IDIOTS 1 :-)

Cheapzilla have been very quiet on the official announcements ( apart from again posting another loss for the quarter ), but there has been a bit of action on the AMD64 front with the release of some Motherboards with the Nvidea Nforce 3 Chipset, which has opened the market up for the Opteron to be utilized for Single Processor Workstation Applications. All Good I hear you say ... well maybe not so good for the AMD64 Hype Wagon I'm afraid with the Opteron Solutions being walloped by both the PIV and even the Athlon XP in a whole bunch of Real World 32 Bit scenarios. Ooouuuchhhh.. Read More Here

Seems like in the short term at least, the whole 64 Bit bubble could burst for AMD with the lack of not only an AMD64 optimized O.S but also any tangible real world 64 bit applications on the immediate horizon. All of the Hot Air and Hyperbole surrounding their 64/32 bit capabilities is proving to be little more than a fart in the wind when the Opteron is being outperformed by their own Athlon XP in most 32 Bit App's.

The Athlon 64 is still being touted for a September release, so hopefully AMD can find some much needed Clock Cycles as well as some more 3rd Party support , and turn it around. As it stands now, they are having enough trouble keeping up with the current PIV-C , let alone trying to contend with the Prescott.

Next ..

 

 

Apple have been a little quieter this month, ( no doubt they are preparing to whip the disciples into a frenzy when the time is nigh) but that hasn't stemmed the fallout from the G5 Paper Launch. The debate over the Benchmarks has raged unabated with both sides of the fence, as expected, finding their own truth's ? in the vast variety of techno babble, to prove their respective arguments for or against.

There have been some amusing moments when some VP's from Apple and IBM were quizzed on the claims of being the First {adv:Before or above all others in time, order, rank, or importance:} and Fastest {Self Explanatory}.

The VP's were asked what they thought about dual-processor machines based on AMD Opterons, which have been available from BOXX Technologies since
4th June, i.e. before the G5 Paper launch. After some mumbling about the Boxx units being a Server not a Workstation or Desktop System (Wrong again) and some other really unconvincing answers Mr. Chekib Akrout, who is Vice President PowerPC & Networking Technology Development at IBM said that in his mind Opteron is not a 64-bit CPU.

You IDIOT !!!

Why the PowerPC 970 (G5) is considered a 64-bit CPU, while Opteron is not, especially in the light of the AMD and IBM cooperation ( Hyper Transport ) is beyond me.

Please, can someone throw me a bone here :-)

Read the whole interview Here

The Apple One Eyed fraternity as expected have also come out guns blazing to defend the almighty Jobs, and strike down anyone and everyone who dare question the validity of the independent ? VariTest results. While some rational and reasonable debate has been conducted , the vast majority has descended into nothing short of blind fanaticism, and at times has been quite comical with accusations of :quote: " if the straight facts tend to be inconvenient to your position, then bury the opposition under a mountain of meaningless techno-babble."

Classic Case of the Pot Calling the Kettle Black don't you think !!!!!

Apple tried burying "the straight facts" so deep in techno babble, that unfortunately for them, the sheer weight of bullshite that they tried to pile on top to hide what was actually going on, was too much for the X86 Evangelists to ignore, and they had a relatively easy task of picking thru the mess.

The whole Argument that Intel also " load" the Spec Test by using optimized compilers doesn't really hold much water when Apple themselves used optimized compilers to Spec bench the G5, but conveniently changed the rules for the X86 units.

The whole point of SPEC is it provides a common repository for purposes of comparison. The fact is Apple deliberately ignored the true SPEC scores for x86 (which handily defeat the G5 scores) to create a misleading impression.

What Apple showed was a benchmark of GCC performance, not Integer or FP performance. And it's no surprise that Apple won, considering GCC is it's primary compiler, whereas serious developers on x86 use compilers from Microsoft or Intel.

As for the other benchmarks, well I had my say last month, in short Apple had results in the past showing the G4's wiping P4's on Photoshop when most non-Apple benchmarks showed the x86's with a significant lead, so that ones up for grabs when the G5 actually ships. Can't wait for that one ;-)

The Audio bench mark lost all credibility when Apple
changed the results from Tracks to PlugIns. They still have 2 Conflicting versions posted on their own web page, one in the White Paper and one on the performance page.

So what is it ; More Tracks or More PlugIns ?

Considering track count has more to do with Hard Drive thru put, that one really didn't make any sense to me, and then to use the exact number for PlugIns is just comical. It just doesn't make any sense .Either one doesn't hold up to scrutiny anyway, and after watching the video presentation again, I smell an even bigger Rat!

IMO: The way the CPU load shot up from around 25% at idle to overloading smacks of de-normaling to me. Nice One Boys. Lets rig the session with a badly coded PlugIn that causes de-normaling in the PIV, and let the thing fall over in front of the dearly devoted who wouldn't know their arse from their elbow to what was really going on. Nice Try Batman, But No Cigar :-)

Even a leading Mac Audio orientated site Here has questioned the merit and validity of Apples attempts at "Preaching to the Choir ", and has posted a scathing attack on the so called " Comparitive Benchmark" !

By the time Apple actually ships the units, their claim of being the first 64 Bit Desktops will definitely be cut to shreds with multiple instances of both Dual and Single processor Opteron Systems being available to the market weeks and months ahead, and as for being the fastest, well lets hope for the one eyed's sake that the units can hold up to Apples lofty claims..Hmmmm
I somehow doubt it
, and with Opteron and the New Xeon DP to contend with already, not to mention the Prescott, I 'd say the Mac Fraternity had better pray to Jobs that the wet dream doesn't turn into a nightmare .. :-)

Interesting Times Ahead..

Switch :-)


© TECHNOLOGY BYTES 2005

Get a cool web address! Free SubDomains from ShortURL.com