My retail Parhelia 128MB has significantly lower core/mem speed.

[Haig]

@Haig: does that depend on the part not supporting a higher speed and hence being factory set at that speed or is it a "bug"?

It's not a bug. We're pretty strict on stability so all tests are done at clocks higher than what's advertized. Whatever passes our tests at the higher speeds get clocked at the advertized speed. If a chip fails, then we do another test at 1% less. If it passes at this speed, then we clock it 1% less than what's advertized. This goes on until a chip doesn't pass within the 10% margin that we advertize.

Hoping that your customers will be ignorant enough not to check the clock speeds and including some fine print saying "your clocks might actually be 20% lower than someone else's" is NOT the way to make customers happy.

Hopefully, we'll go back to our previous ways of not advertizing any clock speeds. P was the first time we did this and personally, I hope it's the last.

[bsdgeek]

How many issues like this are Matroxusers going to have to stand for, hmm? Banding, Clock speeds, lack of proper drivers and feature support, deceptive advertising with hidden cryptic messages, and probably some new issue will pop up soon enough. Some of these I might be able to take, but they just keep adding up.

What is the deal with this? Matrox doesn't have a giant userbase, yet almost all decisions made there seem to piss us off. Tell me, is your management trying to lose customers?

You hope that Matrox goes back to its old ways, eh? I could be way off here, but I think it's mentalities like this one which have relegated Matrox into a niche player from market leader. Matrox has never changed with the times, and it shows.

Note: Haig, if it seems like it in any way, I'm not trying to attack you personally.

[Haig]

You hope that Matrox goes back to its old ways, eh?

Regarding advertizing clock speeds, yes I do.

I could be way off here, but I think it's mentalities like this one which have relegated Matrox into a niche player from market leader.

I guess I'll no longer be nomiated for pres

Note: Haig, if it seems like it in any way, I'm not trying to attack you personally.

Well, you do seem to be totally against my personal opinion.

[bsdgeek]

Originally posted by Haig
Well, you do seem to be totally against my personal opinion.

People can still disagree, but not attack each other.

[Haig]

Yep

[Wombat]

Anybody have a Parhelia retail box lying around? Does it actually have the clock speed on the box?

[UtwigMU]

I heard price difference for dealer between retail and bulk is 20 Euros - 5% diffrence and they order directly from Matrox in Ireland.

While I think selling bellow advertised clock is problematic, I see little point in doing so.

[Chucky Cheese]

Parhelia Box

Wombat,

I have a brand new P still in the box(..retail), which has never been installed. I looked the box over and I do not see the core/mem speed stated. There is Horiz/Vert refresh rates listed on one of the side panels for different resolutions.

let me know if there is any other info you need from the box.

cc

[KintaroOE]

one question....

how fast Clocks the Matrox Parhelia HR256 ???

My old Parhelia, died few days ago. Now i pay 110 € more and i can buy a HR256...

but, i need to know how fast it Clocks...

thanks

[Kurt]

Originally posted by Indiana
No. I'm basing this on the fact that Matrox officially states this on their website.
This is thus NOT an "assumption", but according to Matrox' own website and Haigs post "normal" behaviour.

And there were other reports about surprisingly low clocked retail Parhelia cards not only on this forum, e.g.:

http://forums.murc.ws/showthread.php?s=&threadid=40926&highlight=Parhelia+AND+clock

Besides, I'd really like to see a +10% clocked Parhelia directly from Matrox, since even oc'ers only managed about +25% for Core and mem using crazy coolings and voltage-modding that will most likely kill your gfx-card in a very short time....

Sorry but I haven't found overwhelming evidence of M dumping defective chips on the market.

If you're right then the problem will get clearer with time as more ppl report lower clocked products.

Just don't jump the gun

[UtwigMU]

Originally posted by KintaroOE
one question....

how fast Clocks the Matrox Parhelia HR256 ???

My old Parhelia, died few days ago. Now i pay 110 € more and i can buy a HR256...

but, i need to know how fast it Clocks...

thanks

Parhelia 256HR costs 2500$, unless you have good sources, it's not going to be 110€ difference. It's also PCI 64-bit/66MHz and not AGP card.

Vanilla 256MB AGP is clocked @ bulk speed (200/250)

[Tomasz]

In my opinion, the cards should be tested in the following manner:
Test at 220 MHz (or 225 to give some headroom): If it passes, clock it as a retail card 220 MHz; if it fails then put it in the oem bin.

Retail cards, especially $350-400 retail cards, should not have such large variances in clockspeeds. If I spend that much money on a video card, I would at least want to know beforehand exactly what I will get.

[bsdgeek]

Originally posted by UtwigMU
Parhelia 256HR costs 2500$, unless you have good sources, it's not going to be 110€ difference. It's also PCI 64-bit/66MHz and not AGP card.

Vanilla 256MB AGP is clocked @ bulk speed (200/250)

But remember the 10% variation Utwig, which means you could get a card running at 180/225.

[Chucky Cheese]

HR256???

KintaroOE

My old Parhelia, died few days ago. Now i pay 110 € more and i can buy a HR256...

if you can afford the HR256, why worry about that! This card has an MSRP of US $2495 and is designed to power some very expensive displays...

http://www.matrox.com/mga/products/p_hr256/home.cfm

cc

[a_h]

Haig, may you clarify the testing conditions a bit?

Since an official "fix" exists for the 217 MHz cards, but as far as I understand that these chips didn't achieve satisfactory results at 220, I would be interested how hard the testings are.

Thanks and greets from Austria, Hannes