Pitou

[Voyager3D]

Originally posted by Seajnky
if mb didnt have pro socket couldnt user just use up first pci slot? Small sacrifice for much faster video card.

Maby you have somthing there...
I say a new Parhelia had a cooling solution similar to Geforce FX that used up PCI slot 1, why not take the extra juice from that PCI slot...

[DGhost]

Originally posted by Voyager3D


Maby you have somthing there...
I say a new Parhelia had a cooling solution similar to Geforce FX that used up PCI slot 1, why not take the extra juice from that PCI slot...

you can't guarantee if there is a PCI slot right next to the AGP slot or not. nor can you guarantee that the PCI slot would provide enough power. It's wiser to either use AGP Pro (which limits the market down) or do what 3Dfx, ATI and NVidia did, and put an external power connector on it...

[Reckless]

I think the 'use up PCI slot 1' simply means use the space as nVidia have shown with their monstrosity of a cooler. Yeah, it's technically able but do we seriously see this as the way forward

[DGhost]

honestly, i don't see it as a step back. with most computer systems you don't want to use PCI 1 when using an AGP graphics card, and to boot it helps reduce overheating and allows it to work in a variety of systems. sure its a bit bulky, but compare that to the number of people i've see that have a Voodoo5 and a squirrel cage fan exausting next to it. or the amount of people who mod their gfx cards and put huge hs/f's on them that block that slot anyways. this works better, because not only is it cooling the core, its exhausting it.

Come on, look at how many people modded their Parhelia.

oh, and btw i was responding to the "why not take the extra juice from that PCI slot... " part of that...

[Reckless]

My Asus CUSL2-C mobo has slot 1 PCI with a higher rate for PCI interrupts. It holds my Adaptec 29160 SCSI card and so is well suited to the slot it's in. Also, IRQs at a hardware level are managed on a slot basis and this can reduce the number down. ACPI, will take care of sharing in Windows but there is still contention on the mobo to be aware off. This is probably nano seconds but putting cards in the best slots by examing the sharing table generally lends itself to a more stable system.

I feel the only reason people modded their P's is that it's performance was (and probably still is) mediocre /me runs off

[DGhost]

Originally posted by Reckless
My Asus CUSL2-C mobo has slot 1 PCI with a higher rate for PCI interrupts. It holds my Adaptec 29160 SCSI card and so is well suited to the slot it's in. Also, IRQs at a hardware level are managed on a slot basis and this can reduce the number down. ACPI, will take care of sharing in Windows but there is still contention on the mobo to be aware off. This is probably nano seconds but putting cards in the best slots by examing the sharing table generally lends itself to a more stable system.

A lot of motherboards have PCI1 and the AGP slot sharing an IRQ. to use a device in both of them is bad. like i said, most computer systems will not want to do it anyways because this can lead to numerous stability problems.

IRQ's are not managed on a slot level, they are managed on an interrupt level. Each slot is connected to an IRQ, and often times several slots wind up having to share the same IRQ's. ACPI does not do anything to resolve this problem, as it forces all drivers to talk to the same IRQ.

It is correct that you don't want devices sharing IRQ's.

[Wombat]

It would be kind of bad to just start taking up that second slot. PCI-Express IS coming soon, and the motherboard could physically look the same, but function very differently under the hood.

[Pace]

I thought the AGP Pro spec covered PCI 1? So, make a Pro card. Oh wait, then Matrox need to learn the definition of quality.

[ElDonAntonio]

Well now that it seems like Pitou is canned, would you care to share what you know and what you were supposed to tell us almost 1 month ago Vigilant? Or have you already said everything in your rant?

Thanks man.

[Amiga Blitter]

HA HA

[gordo-uk]

I don’t recall seeing any real evidence that pitou ever existed, some ‘susceptible’ people went OTT as usual, others wanted more concrete evidence and continued to press for some, so here we are a month later still without any details and low and behold!…………….its now been canned, Really.

Did pitou ever exist vigilant?, if it did, you made a serious error of judgement talking about it until you had concrete evidence from your ‘source’ that matrox would go ahead with it. If it did not exist, your reputation has just taken a serious nose dive.

This is NOT a personal attack on you vigilant, so please do not take it as such but this kind of rumour spreading is not helpful to matrox as it tries to justify the Parhelia to the ATI and Nvidia gaming public.

[Enak]

Urm, I seem to have missed the post that P2 has been canned...

[VigilAnt]

Originally posted by gordo-uk
I don’t recall seeing any real evidence that pitou ever existed, some ‘susceptible’ people went OTT as usual, others wanted more concrete evidence and continued to press for some, so here we are a month later still without any details and low and behold!…………….its now been canned, Really.

Did pitou ever exist vigilant?, if it did, you made a serious error of judgement talking about it until you had concrete evidence from your ‘source’ that matrox would go ahead with it. If it did not exist, your reputation has just taken a serious nose dive.

This is NOT a personal attack on you vigilant, so please do not take it as such but this kind of rumour spreading is not helpful to matrox as it tries to justify the Parhelia to the ATI and Nvidia gaming public.

I am not here to gain reputation or credibility. I am here to offer news on a company that is not very open to the public. I also enjoy commenting on their state and do editorials.

I am not here to please you, but I am here to offer information.

Pitou is taped out. It is a project of the follow up of the Parhelia. It did just recently get put on the shelf because of other issues. MGI is going through some troubles and it remains to be seen if they will be able to work themselves out of it. This is one of the reasons I have held out on letting out all I know about the chip, which is a lot more than you think and a lot more than what Matrox would want me to know . The fact is that Parhelia as a project was put on the shelf a while ago and was dusted off again, but had features added and deleted. I was confident that about a month ago Matrox was going to move forward with the project and that would set in stone without doubt the general direction of the chip. Now that it is collecting dust, it is absolutely possible that Pitou will undergo more changes depending on how far it is pushed back IF we see it at all.

VigilAnt

[Dogbert]

Seems like another G800 to me

[VigilAnt]

Originally posted by Dogbert
Seems like another G800 to me

No, G800 was mishandled project from the very beginning. It was actually a play on what the G450 was going to be but ended up missing out because of the time-frame it was introduced. What is interesting is that the G550 core does have features disabled that, had a proper revision been done, would have put it in a much higher class.

Pitou's situation is much different. The chip is all ready to move to the next stage, but this will cost money (and lots of it since they don't have the facilities other chip makers do) and the owners are pulling the reigns in on expenses (witness the layoffs).

Speaking of layoffs, more are to happen this Friday it seems. This round is going to be large.

VigilAnt