![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
-
Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.
When will AGP vanish?
- Thread starter lordsegan
- Start date
Elledan said:
Since the BTX standard cases haven't started shipping yet and no BTX motherboards will be available until sometime early next year, you're dead wrong. BTX is a new standard. It's not dead as it hasn't even started shipping yet. The only BTX case you can get now is the CoolerMaster Stacker case that's convertible between ATX and BTX.
Just what is the basis for your statement?
Moloch said:Its pci-express only.
Nvidia hardly has any pci-express cards(high end) and ati isn't doin agp cards anymore
Nvidia has 6800GT's shipping in PCI-E. Ultra's are do out very soon. So I wouldn't say they hardly have any PCI-E cards. Thier lineup is missing the PCI-E version of the Ultra and the UE is vaporware anyway so it doesn't matter.
ATi's the one who's not doing AGP it seems anymore. As the bulk of thier newer cards are PCI-E only.
I just sold my 6800GT, should have just held onto that a while longer, couldn't stand that loud fan though (I have a shuttle, single slot cooling only).Moloch said:Its pci-express only.
Nvidia hardly has any pci-express cards(high end) and ati isn't doin agp cards anymore
It's difficult to justify spending $500+ on a card on the way out, even if it will be a year or two before it's totally "old school". The biggest "duh" in the whole situation is that the seller is ditching the X800XT PE to move to PCIe, ha ha.
GlobalFear
2[H]4U
- Joined
- Nov 22, 2003
- Messages
- 3,631
Correct me if I'm wrong but, we have not even come close to turning agp into a bottleneck in systems. The only real benifit is SLI, And since ati doesn't have sli (atleast right now) other than to keep the market hot there is no reason for ati to jump on the bandwagon. 3 years before agp is gone. Then again look at most of the pcie boards... Alot of them still have pci slots.
GlobalFear
2[H]4U
- Joined
- Nov 22, 2003
- Messages
- 3,631
Sir-Fragalot said:
Badger_sly
[H]ard|Gawd
- Joined
- Mar 20, 2002
- Messages
- 1,594
For 2+ years AGP will be an easy option to find. 2.5 - 3 years from now, AGP will become slightly more difficult to find. In 3.5 - 4 years AGP will become real legacy hardware, similar to what happened to ISA.
The reason AGP won't die off overnight is that there is hardley any advantage to going PCI-E, and, the average computer owners (who are the real majority, not us tech-heads) don't buy whole new systems more often than 3-5 years. Basically, AGP will be weaned out as the majority of consumers upgrade their computers over the years.
The reason AGP won't die off overnight is that there is hardley any advantage to going PCI-E, and, the average computer owners (who are the real majority, not us tech-heads) don't buy whole new systems more often than 3-5 years. Basically, AGP will be weaned out as the majority of consumers upgrade their computers over the years.
Badger_sly
[H]ard|Gawd
- Joined
- Mar 20, 2002
- Messages
- 1,594
I(illa Bee said:
Again, it'll be dependant on the majority, the "agerage joe". And that is not us that know about PCI-E and SLI. Even for gamers, the number spending high dollars to go SLI and buy 2 expensive video cards, is even less.
Badger_sly said:
It won't die overnight. But don't expect it to stick around that long. Manufacturers want to make the transistion for reasons I stated above. We can already see that happening. It won't hang around in the mainstream more than 2 years.
GlobalFear said:
Except for the video card, the other components don't need that much cooling. With the CPU being cooled so much better, ambient case temperatures will drop significantly. Overall components will become cooler. Heat rises and the video card is now opposed component side up. With proper exhaust fans I think video card cooling will improve also. BTX doesn't get rid of your case fans. I think the rest of the system cooling will be about the same.
Some computers from OEM's aren't ATX standard as it is. Redesigning won't be required for those systems anyway. At the rate OEM's redesign cases and systems the BTX changeover won't be that big a deal. Where it will hurt the most is where it causes divergance between AMD and Intel based systems. OEM's can't use the same chassis for both platforms anymore.
AMD won't be making the change anytime soon to BTX if ever. If anything hurts the BTX standard, that will. But I think AMD will change over eventually.
Badger_sly
[H]ard|Gawd
- Joined
- Mar 20, 2002
- Messages
- 1,594
What "some" manufacturers want, and what will really happen, will be two different things for the most part.Sir-Fragalot said:
And, what is happening with PCI-E and SLI coming out is minuscule in the whole market. Because you or anyone else in this forum has or has ordered a PCI-E video card, and has or wants SLI, doesn't make AGP die any faster. It's understandable that because you spent money on the new technology, you'd want it to become more mainstream quickly, but that doesn't mean it's going to happen.
.........
Badger_sly said:
Ummm no. I work on alot of machines in the retail market. Alot of machines that are comming out and are being sold TODAY are PCI-Express based. Go to your local Comp USA and Best Buy. Most of the Compas and HP machines are PCI-E. They have been for awhile. AGP/PCI models are being phased out.
Intel made a big push for LGA775 and PCI-Express and after more than 6 months, manufacturers are bringing these to market in retail machines. I do alot of memory and video card upgrades for new machines that are purchased, and I see alot of PCI-Express stuff.
What I've ordered for my machine has nothing to do with it.
There is still a 98% base of machines or something that are AGP/PCI. Regaurdless of platform. That's because those are already in place. Go to manufacturers websites and configure a machine there, go to a retail store and start looking at machines. You'll be surprised how many PCI-E systems are out there now. In one-two years (at most) every machine sold will be PCI-E. It replaces both standards, unlike AGP which just added the graphics port. PCI-E replaces the entire I/O system on the the motherboard.
The base of new machines for PCI-E is a much larger percentage than you think. AGP video cards will continue to be made for the low to midrange past that, but AGP as a high end thing is pretty much a thing of the past. Nvidia may do one more round of high end AGP cards, but it doesn't look like ATi's going to.
Are there any chipsets forthcoming that support AGP? I'm thinking all the new chipsets on the roadmap are PCI-E only. I could be wrong. I think that will have more to do with it than anything else. If motherboard design moving forward either is based on existing chipsets to support AGP or new chipsets to support PCI-E, then take inventory now, 'cause there won't be any new AGP motherboards after the current chipsets have all been played out.
rcolbert said:
SIS has released one recently for P4 or it's upcoming. Can't remember which. Also I think VIA has one in the pipe for Pentium 4. Intel will not do anymore AGP/PCI chipsets.
Other than the one that either just came out from SIS or will be comming out and the VIA one, that's it. No more after that.
I strongly disagree with this statement. The chipset (NB and SB), voltage regulators (those things can get very, very hot) and other components like RAM require a lot of cooling.Sir-Fragalot said:
Completely false. What happens is that all of the air, freshly heated by the CPU's heat, is blown over the rest of the components inside the case.
BTX is the complete opposite of ATX in this regard: with BTX the CPU is cooled first, then the rest of the components. With ATX the CPU is right next to the exhaust, allowing its heat to be exhausted immediately (ducts!).
As you might know, the warmer air gets, the less heat it can absorb. In order to remove as much heat from a system as possible it's therefore most efficient to move the air over the coolest components first, to end with the warmest components. This is where ATX gets it right and BTX totally wrong.
The only good part about the BTX spec, to be honest.
No, it won't, because the videocard is being scorched by the 100+ Watt hairdryer at the front of the case (Prescott CPU + thermal module).
This thermal image of a desktop BTX system I've got shows that BTX has the lousiest cooling solution ever devised. I expect lots of premature deaths of components inside BTX systems, as well as many unstable systems.
BTX is unsuitable for dual CPU (NUMA) designs, like the Opteron, so that's at least one market where BTX will never become a success.
Scheizekopf
[H]ard|Gawd
- Joined
- Jan 31, 2003
- Messages
- 1,694
I wont switch until I see the need to. (Performance/Money ratio is what I mean)
Elledan said:
I had this conversation in another thread. First off, I realize that less fans are needed. However manufactures can use as many fans in thier case designs as they wish.
In the BTX case I saw, the video card didn't sit behind the processor. So as the heat is exhausted through the case it won't hit the video card. It will hit the motherboard, but I don't think it will be any worse than it is now. Since air flow through a system isn't very good in relation to motherboard cooling right now anyway.
Right now I am looking at the BTX layout. I see that in all likely hood the air that comes off the CPU thermal solution will pass over the North and Southbridge chips and out the back of the case. The way I see it, the air that moves over it won't be as hot as you might think. And if evacuated from the case fast enough it shouldn't be a problem. Periphrial card cooling looks like alot of the heat that rises from the video card will exit the case where the CPU exhaust does.
I see what your saying and I disagree with you. For one additional reason. The engineers at Intel that are designing the BTX form factor, are smarter than either of is. They got the job designing this crap and we didn't. Sometimes because of cost considerations and space considerations you have to give and take with the design and decide where your priorities lie. I think that although BTX will have some disadvantages, overall I think it will be an improvement.
Before this releases, it has to have been tested. If it were so bad and so much worse than ATX it never would come out.
We aren't going to agree here and that's fine. If it's bad I think the OEM's would have cought on by now and said no way are we going to adopt this. But they haven't. So like it or not it's comming and it might now last, but everyone said the same thing about ATX too. For different reasons of course.
I do agree that BTX won't work very well if at all for SMP systems. Unless they have a EBTX or something that creates the same basic form factor, but with a second CPU cooling solution. This of course would make for a very large case though.
This is a very salient point. Will there be any new AGP cards made?Sir-Fragalot said:
I'd guess yes, but not immediately. Look what some mobo makers have done to incorporate the 865/876 chipset with 775. I doubt Intel had that in mind.
The point being is that there are millions of people with AGP boards full of RAM and a processor, and many of them don't want (or can't afford) to have to replace their entire system just to upgrade a video card. So there will be a market for AGP cards for the foreseeable future.
I think you're right. I own a Socket A Shuttle right now. I am looking at the SN95G5 Shuttle that was recently released. It is Socket 939 (the latest mobo/processor) but they chose to leave AGP video which confuses me a bit. I like my Shuttle and am leaning towards sticking with one. A new processor is all I should need (as long as my ram is compatible) and I'm good to go. If I consider PCI-E it's a whole new level of price vs. performance.uclajd said:
uclajd said:
And for those reasons AGP cards will continue to be made. But I think it's days as a high end part are pretty much over with. Expect maybe one more generation to be offered in both forms and then after that, AGP will go the wayside.
On the motherboard side, SIS just has or is going to release a AGP/PCI Pentium 4 chipset. I imagine this will be one of if not the last ones for the Pentium 4. Intel themselves has over 70% of their own processors chipset market. Therefore since Intel isn't making AGP/PCI chipsets anymore, PCI-E will certainly replace AGP for the Intel processor based PC's.
Intel is still providing legacy chipsets to third party board makers for boards like the P4P500 and the AS8, but don't expect it to continue very long. Currently Intel is reporting shortages of those chipsets. Once their stock of i865's and i875P parts are depleted I expect them to cut their i915x and i925x series chipsets prices down to entice motherobard makers to adopt those chipsets even faster.
Intel also announced a cheaper lower performance chipset for OEM's and board makers to use that cuts costs down vs. the i915's that are out now. I915's are still pretty expensive as far as chipsets go right now. So that should come as good news to board makers.
I agree that the AGP cards in their present models will be available for some time. IMO the best AGP cards available now move the bottleneck off the video card and onto other areas of the system. Having said that, I don't know that there's enough of a market moving forward to justify continuing to release newer AGP cards as simply upgrade parts for existing systems. The R&D expense for ATI and Nvidia to develop new cards on both platforms will be prohbitive, not to mention any engineering tradeoffs that would have to occur. ATI's already clearly on this path by releasing several PCI-E only products (or at least announcing them which for ATI is synonymous with releasing them.)
I do expect that 3rd parties will continue to offer AGP cards based on the design of PCI-E only cards, similar to how 3rd parties provide non-standard features on older cards as well (such as dual-DVI on the Radeon 9600XT.)
Long story short is that we're approaching the freezing point for AGP technology.
I do expect that 3rd parties will continue to offer AGP cards based on the design of PCI-E only cards, similar to how 3rd parties provide non-standard features on older cards as well (such as dual-DVI on the Radeon 9600XT.)
Long story short is that we're approaching the freezing point for AGP technology.
Badger_sly
[H]ard|Gawd
- Joined
- Mar 20, 2002
- Messages
- 1,594
CRXican said:
It depends on how much of a value you put on the big performance gain between a 9800p and a 6800GT, that you'd gain for a few years.
With BTX, the expansion cards will be sitting in what basically amounts to a dead air pocket. The air coming from the thermal module will be exhausted by one or more exhaust fans at the back, meanwhile warming up the air around the expansion cards.Sir-Fragalot said:
In other words, BTX makes the airflow through the case completely ineffective in cooling anything but the CPU.
At 80-100 Watt for the CPU? Aside from the GPU, this is more heat than produced by the rest of the components together.
Which would mean more, and bigger fans. I thought we were trying to improve on ATX, which could do with only a couple of fans at most?
Again, that's the part of BTX which I actually like.
Are we talking about the same Intel which recently cancelled the P4 line and the LCoS project? I still maintain that the BTX formfactor was created to compensate for the obscene power usage of the Prescott core.
Methinks their priorities lie with covering up the blunder which is the Netburst architecture.
(Baby-)AT and all other formfactors before ATX lacked certain features which ATX introduced. BTX on the other hand doesn't offer any significant features over ATX's.
Hence one of BTX's failings compared to ATX. ATX is simply more versatile than BTX.
Elledan said:
Let me make a quick list of typical expansion slots that give a flying fark about heat....
1.) Videocard.
2.)..... Hrrm.. there is no second item. Okay, all done with list.
Actually BTX cools the videocard far better than ATX (since it blows an immense amount of air directly across the hottest part of the video card), and it also cools the RAM (In part because of the positioning of the Powersupply fan), the Northbridge, and the South Bridge far better than ATX does.
[/QUOTE]At 80-100 Watt for the CPU? Aside from the GPU, this is more heat than produced by the rest of the components together.[/QUOTE]
If their is a large amount of air passing quickly over a large heatsink then the air coming off the heatsink won't actually be heated all that much (try checking how cool air coming directly off your current HSF system feels).
Also the BTX specification causes a good bit of air to flow directly under the motherboard (and increases the space for that air flow), which should help lower overall system temps.
Elledan said:
You may or may not be right on any or all of your points. I won't argue the points anymore. I would be willing to bet that it will work better than you think it will.
arentol said:
Elledan said:
arentol said:
I completeley agree with you. I sit there and look at the BTX layout and can see how it should work better than ATX.
The biggest thing I don't like about the new form factor is the case opening on the wrong side. Like a damn Macintosh. That will take a LONG time to get used to.
Soundcards don't like heat much, either. Nor do RAID cards (real ones, with onboard processor and RAM).arentol said:
And what if you've got an SLI setup? Will one card be cooled properly, while the other one is slowly being fried in its own and the CPU's heat?
With ATX, air flows from the front, over the RAM, HSF, voltage regulators to exhaust. With BTX, air flows from the thermal module, over the NB, over the RAM, to the exhaust. Conclusion, the air flowing over the RAM modules will be a lot warmer with BTX.
And your second videocard (SLI or PCI) still isn't being cooled as well as the first one.
You absolutely do not want air from the HSF to pass over the NB, as the HSF is warmer than the NB's HS, thus you're not cooling the NB properly.
So, we're back to the 'big fans, lots of airflow' solution which BTX tried to avoid, right?
That's like saying that the patient died during surgery, but at least you managed to take out his appendix.
Looking at the BTX Interface Specification v1.0, page 9 shows three examples of BTX layouts. Starting from the left, the expandable tower (side view), a number of flaws become apparent:
- there's no space for HDDs whatsoever. If you were to replace the front side I/O peripherals with space for HDDs, things would be better.
- any power cables and P/SATA/FDD cables passing in front of the thermal module will block airflow.
- the expansion cards aren't cooled at all, except for the first (top) one.
- the thermal module wastes a lot of space, which is actually used with ATX.
The desktop is even worse:
- the optical and HDD drives are in a dead air pocket, with no possibility for cooling.
- the expansion cards (except for the one closest to the exhaust, if there's one) are in a dead air pocket
- again, the thermal module is a gigantic waste of space.
- cables crawling through the case will severely affect airflow.
To be honest, current SFF systems look better to me.
Also, BTX is larger than EATX boards, microBTX is larger than microATX and picoBTX is an absolute joke compared to SFF mainboards and mini-ITX. This makes creating an efficient case layout a difficult task.
Browsing through the specification, you can see that aside from the fact that pretty much everything is defined (compared to the ATX spec), Intel had to come up with some hacks to prevent the thermal module from ripping a part out of the mainboard's PCB: the SRM (page 8, page 19).
So I ask again: aside from the top videocard being cooled properly, on what points of ATX does BTX actually improve?
The problem is that you're not thinking like an engineerSir-Fragalot said:
How, exactly? Cooling isn't better, positioning of components like HDDs is poor at best and a lot of space inside the case is wasted.
Which side the mainboard is mounted to is no intrinsic feature of a formfactor. It's been done with ATX as well (Lian-Li's V1000, etc.).
I'll let the people who designed the thing explain it. I agree with what they are saying.
http://www.intel.com/update/contents/dt10031.htm
Why I think ATX needs to be replaced:
ATX does a poor job bringing air in. The air it brings in is blocked by the drive cages and cables for them. The ram blocks air from reaching the CPU or motherboard components. When the air hits the ram (as it is usually opposed horizontaly) it hits the one chip and very little air will get around that first chip.
Not to mention a standard 80mm fan doesn't get much air over 12 inches in distance with only 32CFM's being pushed. It wouldn't reach the CPU or the bulk of the mainboard components anyway.
Then as you've already agreed, the video card is faced down which allows all the escaping heat to bake the card. Since the video card runs as hot or hotter than the CPU, we can see that this part of the case design needs to change. Easiest way to do that is mount the system backwards so that the card faces up. BTX does this.
Air intake through the front of the case will go where it's needed the most. Right to the CPU directly. This makes sense. Air comming off that heatsink will not be 100 degrees. Right now if I go home and touch my Prescott LGA775 heatsink, its warm, but not too hot to touch. The air comming off of it won't be 100 degrees. With greater CFM's of air comming through and hitting a larger heatsink. It will do nothing but drive the temperature down. If the video card mosfetts can handle the heat of being placed in an area where the heat can not escape, and some video card GPU's get hotter than Prescott, why wouldn't the motherboard be able to handle as much? I believe that mainboard components will be cooled better by the new system. The air comming off the cooler heatsink and from the intake fans will suffice.
The memory doesn't get cooled at all right now as most if not all the air is blocked by a drive cage. The way ram sits as I mentioned before, keeps all but the first chip completely blocked from air in most systems. This is pretty useless.
Ram on BTX is mounted vertcally. Therefore if air flow is directed there, it will go across the memory modules. This has potential in keeping overall system temperatures and memory temps down. Since ram isn't being cooled worth a crap now, it's nothing but an improvement here. Even if it wasn't, its not any worse. Also the ram won't block any airflow either.
Video card cooling will increase noticably. Realize that with hard drives in front of them in virtually every ATX case made in the last 7 years or more, virtually no air actually reaches the video card at all.
Secondly on the video card, it doesn't sit behind the CPU. Therefore the CPU won't be baking it. Where does a second video card go? Well right below the other one. AWAY from the CPU. That's a good thing. The video card faces up and hot air rises. It's all exhausted out of the case the same way ATX does. That's no disadvantage at all. ATX does a good job of that. Provided your case is of decent design.
Drive cooling will be no better or no worse. Manufacturers of cases can simply add a fan to the drive cage if they wish. This will be just like it is today. The way some cases like the CoolerMaster Stacker are designed. The PSU can even get cooling from front intake fans should a mount be placed in front of it.
I am not seeing any disadvantages here. Seems feesable to me. But then again, I am not thinking like an engineer.
Finally the G5 has been out for over a year. It has many design similarities to BTX and the G5 chip gets nearly as hot as Prescott. Hence why they chose to employ water cooling for dual 2.5GHz models. These systems are quite. Even the air cooled ones and have 6800Ultra's in them and many of the other technologies present in modern PC's.
If it works for Apple, why not Intel?
http://www.intel.com/update/contents/dt10031.htm
Why I think ATX needs to be replaced:
ATX does a poor job bringing air in. The air it brings in is blocked by the drive cages and cables for them. The ram blocks air from reaching the CPU or motherboard components. When the air hits the ram (as it is usually opposed horizontaly) it hits the one chip and very little air will get around that first chip.
Not to mention a standard 80mm fan doesn't get much air over 12 inches in distance with only 32CFM's being pushed. It wouldn't reach the CPU or the bulk of the mainboard components anyway.
Then as you've already agreed, the video card is faced down which allows all the escaping heat to bake the card. Since the video card runs as hot or hotter than the CPU, we can see that this part of the case design needs to change. Easiest way to do that is mount the system backwards so that the card faces up. BTX does this.
Air intake through the front of the case will go where it's needed the most. Right to the CPU directly. This makes sense. Air comming off that heatsink will not be 100 degrees. Right now if I go home and touch my Prescott LGA775 heatsink, its warm, but not too hot to touch. The air comming off of it won't be 100 degrees. With greater CFM's of air comming through and hitting a larger heatsink. It will do nothing but drive the temperature down. If the video card mosfetts can handle the heat of being placed in an area where the heat can not escape, and some video card GPU's get hotter than Prescott, why wouldn't the motherboard be able to handle as much? I believe that mainboard components will be cooled better by the new system. The air comming off the cooler heatsink and from the intake fans will suffice.
The memory doesn't get cooled at all right now as most if not all the air is blocked by a drive cage. The way ram sits as I mentioned before, keeps all but the first chip completely blocked from air in most systems. This is pretty useless.
Ram on BTX is mounted vertcally. Therefore if air flow is directed there, it will go across the memory modules. This has potential in keeping overall system temperatures and memory temps down. Since ram isn't being cooled worth a crap now, it's nothing but an improvement here. Even if it wasn't, its not any worse. Also the ram won't block any airflow either.
Video card cooling will increase noticably. Realize that with hard drives in front of them in virtually every ATX case made in the last 7 years or more, virtually no air actually reaches the video card at all.
Secondly on the video card, it doesn't sit behind the CPU. Therefore the CPU won't be baking it. Where does a second video card go? Well right below the other one. AWAY from the CPU. That's a good thing. The video card faces up and hot air rises. It's all exhausted out of the case the same way ATX does. That's no disadvantage at all. ATX does a good job of that. Provided your case is of decent design.
Drive cooling will be no better or no worse. Manufacturers of cases can simply add a fan to the drive cage if they wish. This will be just like it is today. The way some cases like the CoolerMaster Stacker are designed. The PSU can even get cooling from front intake fans should a mount be placed in front of it.
I am not seeing any disadvantages here. Seems feesable to me. But then again, I am not thinking like an engineer.
Finally the G5 has been out for over a year. It has many design similarities to BTX and the G5 chip gets nearly as hot as Prescott. Hence why they chose to employ water cooling for dual 2.5GHz models. These systems are quite. Even the air cooled ones and have 6800Ultra's in them and many of the other technologies present in modern PC's.
If it works for Apple, why not Intel?
Thanks for the propaganda. You really think Intel would offer a 'fair and balanced' overview of the BTX format? They're trying to sell something here.Sir-Fragalot said:
Look at this quote for example: " BTX is also optimized for the newest desktop technologies, including PCI Express and Serial ATA.". Heh.
The ATX spec doesn't define anything regarding airflow. At most it recommends that the PSU is used for exhaust. Implementing a layout which allows for good airflow characteristics is the duty of the case manufacturer/OEM. This is by no means rocket science.
Mainboard manufacturers are free to use the space within Area A (refer to ATX spec sheet rev.2.1, page 17) as they please (except for the back I/O section). Many manufacturers often put the RAM slots in a vertical fashion, allowing the RAM to be cooled through convection.
Most HSFs put the fan at the top of the heatsink. The RAM does not hinder the airflow in this configuration. You also may have noticed that most cases have over 10 cm of empty space above the RAM slots.
There's more to cooling with air than just pushing loads of air through a system. In other words, it's not how much you've got, but how you use it. By avoiding turbulence where necessary, as well as dead air pockets, you only have to push enough air through a system to keep its components below a certain temperature.
You'd be surprised to see how much a simple duct can improve cooling performance.
What about just flipping the ATX mainboard?
Also, there are already cooling solutions for videocards which exhaust the air outside the case. This is similar to the ducting solution for a CPU. This way the heat from the GPU(s) and CPU is directly exhausted without warming up the air inside the case by a significant amount.
No it doesn't. Anyone who has studied this subject knows that the ability of air to 'absorb' heat doesn't decrease in a linear fashion. This is why in order to remove the greatest amount of energy (heat) from a system, it's crucial that the air is led over the components in order of temperature, with the coolest parts first. Because BTX reverses this order, it's less efficient at cooling the system than ATX.
True.
Because the videocard's MOSFETs only have to deal with a couple of voltages at most, at lower tolerances than the mainboard's MOSFETs, which not only have to supply a large array of different voltages, to not just onboard components, but also to expansion cards, fans and many other kinds of peripherals (USB devices, for example). Not to mention that the mainboard's MOSFETs have to deal with far more current, and thus heat up far more quickly.
You may believe this, but do you have any evidence to support this claim?
You want to say, for example, that this thermal image of a BTX desktop system I'm looking at right now (look at those pretty red colours trailing behind the thermal module
) is incorrect or false?You already mentioned this.
Please refrain from repeating unfounded statements.
Hence new, innovative videocard cooling designs like the one I mentioned earlier.
Actually, the first PCIe slot is located pretty much right behind the thermal module. Just look at the BTX layout.
Three words: dead air pocket. Not that it's much better than with ATX, but I doubt that any formfactor will solve this issue as long as expansion cards are kept horizontally on a level plane.
You want to say that BTX is like the PowerMac G5 layout? You might want to take a closer look. Aside from the fact that Intel has made a poor attempt at introducing 'thermal zones' like the PowerMac G5 has, most similarities are superficial. Where are the thermal sensors, for example? What about dual CPU, dual RAM slots? BTX allows for none of those, or the spec is very quiet about it.
- Joined
- Feb 29, 2004
- Messages
- 2,514
Jasonx82 said:Lol yeah, when i go to fix peoples computer i always wonder why they use that connecter to connect it to the PS/2 port maybe they think its faster cuz it was made for a mouse? heheh
I actually use this, because when my computer is under heavy load, the mouse delays and pauses. The PS2 DOESNT do this at all. I cannot STAND the USB mouse. Maybe newer mice are better about this, but my old intellimouse explorer that i'm rocking works better for PS/2.
Elledan said:
Yes Intel's marketting people need to leave technical papers alone. Theres no dispute there.
Elledan said:
True. Early ATX designs didn't have that much to offer over AT in terms of air flow.
Elledan said:
I was speaking of air flow through the front of the case and the way typical ATX cases bring air into the machine. They do a poor job of it.
Elledan said:
I know a duct can. OEM computers have been doing it for years. The NV Silencer 5 also uses this approach with the video card.
Elledan said:
I thought of that too. It would work and I believe theres a Lian-Li case that does this.
Elledan said:
Mentioned above. I also agree, I had a NV Silencer 5 on my 6800GT AGP and was rather fond of it.
Elledan said:
I see what your saying here. I think Intel banks on the components on the motherboard will survive and function well enough despite this. BTX simply does a better job of CPU cooling. Which is the main thing they are after.
Elledan said:
Valid point.
Elledan said:You want to say, for example, that this thermal image of a BTX desktop system I'm looking at right now (look at those pretty red colours trailing behind the thermal module
I've seen it too and again I think Intel feels that the tradeoff is worth it. I see your point and wonder about the validity of the design upon further consideration. We'll have to see when it comes out.
Elledan said:
I think that the air comming off the CPU's heatsink won't be hot enough to affect the card either way. The CPU will be cooler in the BTX design. Even if other stuff may not be. So that air shouldn't heat up the video card too much if at all.
Elledan said:
I agree. And that continues to be a problem. But with the exception of a few select types of cards, few expansion cards aside from video get hot enough to need any kind of cooling.
Elledan said:
There are similarities here. Aside from the CPU. The thermal zoning is different, while the basic theory remains similar. The G5 may do it better. We'll have to see.
Which, as I pointed out, is a failure of the case manufacturer/OEM, not of the ATX formfactor.Sir-Fragalot said:
There are countless ways to improve upon current case designs, few of which have been even explored so far.
Like the V1000 I mentioned in an earlier post?
With the Prescott running at temperatures well over 60 C during load with the stock HSF, Intel probably thought that the 'shotgun'-approach would work best to 'fix' this cooling issue. Until they decided to cancel further developments on the P4 core, that is.
So far. Soundcards, TV-cards, video-processing cards, RAID cards... they're all gaining new features, which require more, bigger and faster ICs and thus end up using more and more energy, losing much of if as heat. We don't even know how these and other cards will evolve, and which types of expansion cards will be introduced in 5-10 years time.
You don't claim that you can predict the future, now do you?
The PowerMac G5 is more than its mainboard. Every single part has been designed to work together with the rest of the system. That's why you can't just refer to the PowerMac G5's mainboard (sorry, logic board
). From the case itself, to the thermal sensors and fans, to the CPU's, it's not a collection of off-the-shelf parts.Honestly, a flipped ATX system (like with those Lian-Li ones) has more in common in regards to general layout with the PowerMac G5's than a BTX system.
Ballz2TheWallz
2[H]4U
- Joined
- Jun 29, 2004
- Messages
- 2,352
damn pci-e requires upgradeing and look at the prices for a agp vs a pci-e card with same spec price is to much, unless some big black man sticks a dick up my ass if i dont switch over(basically agp becomes bottleneck or they stop producing agp all together)then long live AGP