VigilAnt Rant and Some Observations

[DGhost]

I would most certainly not ever recommend it as a professional product for the reasons Ryu stated...

it has some fairly impressive technologies, but its too flawed of a product to be worth recommending. problems like the FAA and the banding probably won't be truely fixed without a new hardware release. Problems like the anisotropic filtering might be enabled in on this product, but its gonna be at such a performance hit that it won't be worth enabling.

I can't believe that Matrox's solution to it not doing well in the consumer market is to just label it a professional card. It has the feature support to be a professional card, but its just too flawed of a product.

I have to wonder if that is Matrox's solution to everything... "well, this product sucks so lets just sell it to businesses"....

[Chief_Inspector]

"I have to wonder if that is Matrox's solution to everything... "well, this product sucks so lets just sell it to businesses"...."

Me thinks that anytime you see a card aimed for the business market, it's becuase the card is broken. No one in their right mind sets off from the begining to design a card just for business in mind. You build to specs, it doesn't meet specs, it doesn't pass production diag, so you end up lowering clocks and give a marketing spin and sell it to the poor shleps in business. Anyhoo, Matrox is finished in the 3D market. Unless these boys start building a proper hardware simulation setup, get people that know how to package a chip properly, get experienced asic designers, etc, they are better off closing the entire asic dep't and buy chips from nv or ati. Then again, what kind of experienced asic man would want to work in a company that has 1 P4 system as their chip simulator?

CI

[GT98]

Originally posted by Chief_Inspector
"I have to wonder if that is Matrox's solution to everything... "well, this product sucks so lets just sell it to businesses"...."

Me thinks that anytime you see a card aimed for the business market, it's becuase the card is broken. No one in their right mind sets off from the begining to design a card just for business in mind. You build to specs, it doesn't meet specs, it doesn't pass production diag, so you end up lowering clocks and give a marketing spin and sell it to the poor shleps in business. Anyhoo, Matrox is finished in the 3D market. Unless these boys start building a proper hardware simulation setup, get people that know how to package a chip properly, get experienced asic designers, etc, they are better off closing the entire asic dep't and buy chips from nv or ati. Then again, what kind of experienced asic man would want to work in a company that has 1 P4 system as their chip simulator?

CI

Whoo Hoo more good news

[UtwigMU]

Read this artcile on nVidia's lagging behind:

http://www.tech-report.com/etc/2002q4/geforce-fx/index.x?pg=1

In order to win company needs to be competitive in all segments. It's not parhelia problems that are hurting Matrox it's the G800 fiasco.

Let's take a look at segments:

- Ultra cheap. Due to integrated chipsets, this part will be obsolete soon (ECS hammer and P4 boards with onboard Xabre, nForce2 and intel extreme graphics)

1999 technology, 0~50$ US online prices
TNT2, Ati Rage and Radeon based

Value
This segment brings in most money
2000 technology, 50-100$ price range

2MX, 4MX, Radeon 7500 based, Matrox G450 - not competitive (belongs in previous segment)

Midrange
This segment is also very strong
2001 - 2002 technology, 100-150$ price range
GeForce 3 or 4 Ti 4200, Radeon 9000 and Radeon 8500 based, Matrox G550 - G800 with GeForce 2/3 class performance would fare much better

Performance
2002 technology, 150-300$ price range
GeForce 4 Ti 4400 - 4600 based, Radeon 9500

High end
2002 and future (DX9) technology. 300$ and above price range
GeForce Ti 4600, Radeon 9700, Matrox Parhelia - again not competitive performance wise

Now imagine H02 next year:

High End ($300+)
Parhelia 2???
Radeon 10000 (R350) ???
GeForceFX

Performance 150-300$
Iterations of above cards, Radeon 9700

Midrange 100-150$
Parhelia8x or original, Radeon 9500, GeForce 4 4400-4600

Value 50-100$
GeForce 4 Ti 4200, GeForce FXMX (lame name IMO), Radeon 9000, Matrox G550 (800)

Matrox should cover all the segments with competitive products. Before you say nVidiot thinking scroll back to 1999: High end G400MAX(TNT2 Ultra, GeForce DDR -nV takeover, Fury MAXX, Vodoo 3 3500), Performance G400 32DH(TNT2, Fury, Vodoo 3 3000), Midrange G400 16SH OEM (TNT2 Vanta, Rage Pro - Ati sucked that year as you can see, S3 Savage IIRC), Value G200 (TNT, Vodoo2)), ultra cheap G100, Millenium II (Vodoo 1, other solutions like no9, S3 Virge).

Since Parhelia is not competitive releasing multimedia card and raising price by 50%/retaining current price and lowering Parhelia price (like AIW) is IMO a wise choice:

- cutting Parhelia 128 to 250$ (suplementing with fixed 8x core)
- releasing Parhelia 64 in 150-200$ range (current core with features that cause issues disabled, retain dual DVIs)
- Releasing Parhelia Marvel 64 @ 250-300$ and Marvel 128 @350-400$ (AIW is 432$)
- Cutting Parhelia 256 to 500$ later (january...., suplementing with 8x core)

- Releasing Parhelia 2MAX (256MB) that will pwn or equal GeForceFX and Radeon 10000 in May-Jully High end part comes out first at 350-450$
- Releasing Parhelia 2 256, 128MB at 250-350
- Releasing Parhelia2 64OEM, 128OEM or cutting Parhelia 8x to 150-200 with disabbled features on PCB
- Lowering original Parhelia 64 OEM to 100-150$
- Follow up with Marvel parts

- Clearing the stocks of G450, G550, G400, G200, G100, m3Ds, Milleniums before Longhorn comes by.

That way like 1999 Matrox is teh win.

[Stringy]

Good ideas....

MAny decent and plausable ideas and methods to bring the M back in the good graces of the Computer industry, or really the enthusiast computer industry...

The P isn't all that bad...
it does alot with it's [comparibly] low clock speed...

The only real thing I think M should think long and hard about is how to release the 8x part without losing their shirt on all the overstocked P's...
As some have mentioned a 64mb version attractively priced with Dual DVI [or just DH like G400] would be a smart way to rid the warehouse of all the overstock and still make some money...

Matrox is far from dead...
For a company held privately, and one that has consistantly been tight lipped we have alot of insight into the inner-workings...
We know:
1) Bugs were known and deemed not to have enough impact to require it be fixed before release..
2) features and cards were cut and/or not released even tho they were/are semi ready..
3) massive lay-offs and cuts deep in the company..
4) there is another part in the pipe, and that it has even Taped out....

been interesting to follow,
keep up the good reporting VigilAnt and the MURC...


Craig

[Maniac]

Sooner or later, the harsh reality will slowly begin to settle in. Matrox simply doesn't have the resources, desire, leadership, or foresight to keep up with giants like ATI and Nvidia. These companies are constantly pushing the envelope with almost revolutionary advances. With games like Stalker set to explode onto the scene, its just a matter of time. Unfortunately either M. will have to settle to be a econo PC oem supplier, or it will be scooped up by a large corporation. Nvidia FX is just the first step...

[brokenwallet]

If the P really flopped on the PC market, Matrox could always do some modifications and release them to the Mac designer wannabes. They all pay quite a lot for their hardware anyway, and don't usually upgrade video cards every 6 months. Add video capture to the P and Matrox is probably considered king in the content creation world.

[Admiral]

Re: Good ideas....

Originally posted by Stringy

.
.
.
The only real thing I think M should think long and hard about is how to release the 8x part without losing their shirt on all the overstocked P's...
.
.
.
4) there is another part in the pipe, and that it has even Taped out....

Drop prices on the current P, 200-300$ and introduce the revised P at 300-350$.

About the pitou being taped out, nVidia had the GF FX taped out since March (reported by Anandtech) but couldn't build it because the 0.13 process wasn't in place at TSMC. Since the P2 is supposed to be a 0.9 process it might hit on similar production issues.

[Seajnky]

i have a fealing matrox makes more money than people think. Im guessing they make there money selling to .gov and large corportions. If they buy in bulk, lower distrobution costs. but what do i know.

[GT98]

Originally posted by Seajnky
i have a fealing matrox makes more money than people think. Im guessing they make there money selling to .gov and large corportions. If they buy in bulk, lower distrobution costs. but what do i know.

Well far as I know the Graphics End of Matrox doesn't make all that much for reasons stated above. What keeps Matrox in the black is its Video busniess, which it owns the market with its RT line of products

[Chief_Inspector]

"Matrox is far from dead... "

Regarding 3D, they are finished, finito, caput. No resources, no proper test equipment, no roadmap, no oems, and a scrooge as an owner, is a recipe for a sinking ship. They can keep on doing 2D cards but that will come to an end. They can go after niche markets but these people buy a card every 10 years and expect to get support for that long.

"For a company held privately, "

Yeah, for a 25 year old company, they did good. Sadly enough, the recent massive layoffs with more to come this month and in February along with a salary and hiring freeze for the past 3 years has killed morale and oddly enough, no head honcho is doing anything about it. Sort a tells you what they think of MGI.

CI

[GT98]

So Chief_Inspector...what your saying is don't expect anything from Matrox Graphics that will be "worth a damn" and theres no hope of them turning around the boat and its time to jump ship? Pretty gloomy if you ask me

[Chief_Inspector]

Sadly yes that's what I'm saying. I do hope that I'm wrong but I don't see how they can compete on the 3D side without the proper people & equipement. Let's face it, they come out with their new design every few years or so and then make spinoffs from that new design for ~ 2 years. Just like the G200 to G250 and G400 to G450 and G550. While the others are making spinoffs that are somewhat of an improved and faster version of that initial chip, matrox does it backwards. Meanwhile, that initial chip is in most cases flawed and hardware features are disabled for whatever reasons. Rather than tweaking it and improving speed, they come out with a slower and less featured card. What I'd like to know is how all this backward thinking even got the go ahead.

CI

[Stringy]

Interesting how you damn it as a ship sunk...
one thing I've learned about the PC industry is anything is possible...
Just look at ATi, Many folks wrote them off....

Matrox has alot going for it.... Mind you they are in troubled times, but things aren't all that bleek...
They still have cards, they do have a good base to work from...
And they keep inovating with features such as Triple-Head, FAA, Giga-Color...

They have the People to do it...
they need the leadership and market foresight to put it altogether...

Whether they will be able to put it off, I don't know....
They certainly have gotten pretty far with all these faults, fix a few and this company will shine.... leave the holes and they'll sink

Just opinions...
Craig

[Chief_Inspector]

Opinions is a good thing

"Just look at ATi, Many folks wrote them off.... "

Yes they did write them off. Once they replaced top management, look what happened. 9700 comes out, for once, the drivers are good off the cd, and stocks went up.

"Matrox has alot going for it.... Mind you they are in troubled times, but things aren't all that bleek... "

Well, when people start leaving without finding another job, that sort a tells you how bleek it is in there

"They have the People to do it... "

Where they need big help is in the R&D dep't, which they don't have ANY experienced people. Maybe other departments still have a few experienced people left but they don't make chips. Besides, I think their total layoff count is ~ 500 people within the year and another 100 that left. These 600 people couldn't be that
hopeless can they? That's half the company that's gone in 1 year.
No doubt, they need to do a quick 360 if they want to be competitive in 3D. Question is, do they want to do this and are they able to do a quick turnaround? Time will tell me thinks.

How about this one. Each time there is some kind of debate here about Matrox not having anything new, the beta testers were the first ones to jump in giving hints about something new coming soon. Where are they in this thread?