The problem is that adding a layer of coloring is that you generally increase thermal resistance due to the extra layer.
Not to mention what adding a layer of friggin' plastic (the case itself) does... Ok, so there is an aluminium 'thermal transfer plate' above the core, but I'd much rather have one without the case...
The purpose of pretty proccessors is marketing. Yet another reason to dislike marketing...:-/
Yes, I do too. But the problem is of course cost. Legal struggles ain't cheap. And it's a good bet that the RIAA is aiming exactly at MP3.com's wallet. -
There have been conflicting reports with regards to this. Some say that there will also be thunderbirds from Austin. Rather likely too, since the lower speed thunderbirds will not need copper tech. Also see my cache bin posts above. -
Ooops, I realized something about the AMD case. From the start, there will be no Durons from Dresden (aka Fab 30) - only thunderbirds. But it's a good guess they'll be stockpiling these copper-intercon. high-speed Duron's for a rainy day. Like the day Intel releases a better Celeron. -
Exactly. And they are not the first to do this: Why do you think the Celeron II (by some named the 'Celermine') is physically identical to the Coppermine, apart from marking and multiplier?
Not a stupid strategy either. As far as I know, large caches are one of the hardest things to get good yields on. When a small part of the cache is non-functional, it is convenient (sp?) to be able to sell the part anyway. Half of the cache is disabled, and you have an instant Celermine. (3/4 is disabled in the AMD case.)
It's that visual basic is well.. a run time compiler basic.
No. Modern versions of VB are complied when you 'make executable', as they call it. Compiled to machine code, mind you, not some bytecode (*cough*like Java*/cough*).
1) It is slow 5) Did I mention that it is so slooooooooow
Well, it sure loads slow, if no other VB apps are running, cause it has to load the runtime library. But once that is loaded, it ain't slow. Sure, it isn't exactly hand-tuned assembly, but it is indeed a compiled language. Yes, complied to machine code, at least in the modern versions.
Although programs may look simpler to read and maintain in VB in reality it is just hiding the grottyness under a high level API.
Isn't that what they call "abstraction"? =)
In a real language you can write a compiler for your language in that language and have the compiler compile itself.(This can be done with gcc) VB can't do this.
Well, duh. MSVC++ can do this too. But can Perl? Can Java? Python? Otherfavoritelanguage?
(Yes, I'm sure you can hack together a complier for some of the above languages in those languages, but the same goes for VB. You only really need string manipulation to write a complier, and they all have that.)
2) It is proprietary 3) It only runs on windows
Yes. Ultra-true. This is the real downside of VB.
Also, try to judge VB for what it is. It's a GUI hack language. Kindof like shell scripting or Perl for GUI's. If someone asks you to write, say, a 3D game in VB, I feel your pain.:-) -
Not much other than I only wish PCI versions were still available;-)
Anandtech's article indicated that the GeForce2 would come out in a PCI version (cause NVIDIA can't stand the thought of 3DFX having a market to itself;-).
From anandtech: And taking a page from 3dfx's book to success, NVIDIA will be offering the GeForce 2 GTS in both AGP and PCI configurations
Another benefit of the die shrink is that the GeForce 2 GTS consumes close to half of the power as the original GeForce, putting it at between 8 - 9W versus the 16W for the GeForce.
because it is fixed at 4 times oversampling for each pixel.
Hmm. Strange. I do belive what you say, and yet from the screenshots I have seen, 3dfx's FSAA in 2x mode looks better that NVIDIA's. Oh well, I guess 3dfx's jittered accumulation method really does give better visuals. And I thought it was just marketing...;-)
Anyway, I don't think the V5 has much of a chance against both the GeForce 2 and the Radeon (sp?) in the marketplace. Oh well, too bad about a once good (and open-source supporting) company.
Somebody who knows for sure if Quake3 supports hardware T&L on the GeForce, please chime in.
Yes. I know for sure. Quake 3 uses the full OpenGL T&L pipeline. Older OpenGL games (notably Quake2) often only use OpenGL as a rasterization API, but Quake 3 uses the full pipeline. That's why Carmack was out lobbying all the hardware makers for better OpenGL driver - since if a card does not have HW T&L, the OpenGL T&L is done in the driver, which needs heavy optimizations.
...call me an EE major (I am one:-) but I belive infrared is ABOVE microwave in frequency. Anyway, most of the heat produced by a CPU is transfered through heat conduction, not radiation. I don't think the actual clock signal would cause much radiation. Clocking a CPU higher may increase heat production, but the frequency of the radiation (if any) is not directly related to the clock frequency.
Anyway, even if it did, the high-amplitude microwaves in an oven are stopped by rather thin sheets of steel. And the amplitude from the CPU must be lower. Remember, your oven draws multiple kW, while I damn hope your CPU keeps below 100W. (Otherwise, you have an Alpha, and I WANT IT!:-) -
Re:I speak for myself when I say...
on
Athlons Sold Out
·
· Score: 1
Would be great... but NVIDIA thought of it first, and crippled the GeForce:
It's ultra-fast with everything EXCEPT anti-aliased lines (as in wireframe models, CAD). When drawing AA lines, it resorts to software, with the result being that it is an order of magnitude slower than even a TNT.:-(
Of course, they did it to sell more Quadra's (which is a slightly higher clocked GeForce, uncrippled, and with 64Mb DDR. *drool*).
I just thought of this... maybe this is why they don't want open drivers... the crippling might be done in software. (No, using Quadra drivers doesn't help - they most probably detect which chip is installed.) -
Hotrail have had an SMP chipset for the Athlon available for some time
Is their chipset really done? I have heard of it many times, but never as anything else than a future product. A PC chipset isn't exactly an easy thing to make... hey, even Intel screws up at times.
-
Re:I speak for myself when I say...
on
Athlons Sold Out
·
· Score: 1
All of the Geforce issues have been addressed (in some fashion).
Indeed. And anyway, which sane business buyer would put a GeForce in the machines...? -
Re:AMD is just clearing out old stock
on
Athlons Sold Out
·
· Score: 1
I think he mean "replaced in manufacturing" when he said "replacing".
They have to make them before thay can sell them.:-)
-
Re:AMD is just clearing out old stock
on
Athlons Sold Out
·
· Score: 1
The Spitfire *is* AMD's low end chip... it's meant to compete with the Celeries( hence the Socket-A only package )
And anyway, some tests say that even the Spitfire is faster that todays Athlon, escpecially in apps that don't demand large cache (quake3:-).
Another reason to dump the current athlon is costs. Even if the fabs have to be tweaked to yield good on-die caches, I would say it'll still be cheaper that buying seperate cache chips from some other company.
The actual FPU is identical to a P2, and the x87 part is identical to P3 as well (but no SSE in Celeron 1).
The FPU might benchmark as slower because of differences in cache configuration. The Athlon has a stronger FPU than any of the Intels, but the caches are slower.
Anyway, when the poster you replied to talked about a slow FPU, he was probably comparing it to say an Alpha... *drool*
Bull. AGP is a direct connection to the chipset northbridge. Also: PCI is a 32-bit 33Mhz bus... AGP is a 64-bit 66Mhz, double or quad pumped if 2x or 4x. I don't know what the use of that bus would be if you had to transfer the data over PCI... -
Slashdot will republish and resell your posts without asking for your permission.
But, in a manner of speaking, httpd does this every time someone clicks 'reload' (well, ok, not resell, but for ad profit), and I hope you agreed to that... -
Slashdot Mirror
Black is the color which has the most efficient thermal dissipation due to radiated heat.
;-)
:-/
Exactly. Black-body radiator. Physics. Someone finally remembered.
The problem is that adding a layer of coloring is that you generally increase thermal resistance due to the extra layer.
Not to mention what adding a layer of friggin' plastic (the case itself) does... Ok, so there is an aluminium 'thermal transfer plate' above the core, but I'd much rather have one without the case...
The purpose of pretty proccessors is marketing. Yet another reason to dislike marketing...
-
Yes, I do too. But the problem is of course cost. Legal struggles ain't cheap. And it's a good bet that the RIAA is aiming exactly at MP3.com's wallet.
-
There have been conflicting reports with regards to this. Some say that there will also be thunderbirds from Austin. Rather likely too, since the lower speed thunderbirds will not need copper tech. Also see my cache bin posts above.
-
Ooops, I realized something about the AMD case. From the start, there will be no Durons from Dresden (aka Fab 30) - only thunderbirds. But it's a good guess they'll be stockpiling these copper-intercon. high-speed Duron's for a rainy day. Like the day Intel releases a better Celeron.
-
Exactly. And they are not the first to do this: Why do you think the Celeron II (by some named the 'Celermine') is physically identical to the Coppermine, apart from marking and multiplier?
Not a stupid strategy either. As far as I know, large caches are one of the hardest things to get good yields on. When a small part of the cache is non-functional, it is convenient (sp?) to be able to sell the part anyway. Half of the cache is disabled, and you have an instant Celermine. (3/4 is disabled in the AMD case.)
dufke
-
It's that visual basic is well.. a run time compiler basic.
No. Modern versions of VB are complied when you 'make executable', as they call it. Compiled to machine code, mind you, not some bytecode (*cough*like Java*/cough*).
-
1) It is slow
:-)
5) Did I mention that it is so slooooooooow
Well, it sure loads slow, if no other VB apps are running, cause it has to load the runtime library. But once that is loaded, it ain't slow. Sure, it isn't exactly hand-tuned assembly, but it is indeed a compiled language. Yes, complied to machine code, at least in the modern versions.
Although programs may look simpler to read and maintain in VB in reality it is just hiding the grottyness under a high level API.
Isn't that what they call "abstraction"? =)
In a real language you can write a compiler for your language in that language and have the compiler compile itself.(This can be done with gcc) VB can't do this.
Well, duh. MSVC++ can do this too. But can Perl? Can Java? Python? Otherfavoritelanguage?
(Yes, I'm sure you can hack together a complier for some of the above languages in those languages, but the same goes for VB. You only really need string manipulation to write a complier, and they all have that.)
2) It is proprietary
3) It only runs on windows
Yes. Ultra-true. This is the real downside of VB.
Also, try to judge VB for what it is. It's a GUI hack language. Kindof like shell scripting or Perl for GUI's. If someone asks you to write, say, a 3D game in VB, I feel your pain.
-
Exactly... aren't these disclaimers are found on all sold software, including OS'es, including Linux distros...?
-
Not much other than I only wish PCI versions were still available ;-)
;-).
Anandtech's article indicated that the GeForce2 would come out in a PCI version (cause NVIDIA can't stand the thought of 3DFX having a market to itself
From anandtech:
And taking a page from 3dfx's book to success, NVIDIA will be offering the GeForce 2 GTS in both AGP and PCI configurations
-
From Anandtech:
:-)
Another benefit of the die shrink is that the GeForce 2 GTS consumes close to half of the power as the original GeForce, putting it at between 8 - 9W versus the 16W for the GeForce.
Need I say more...?
-
because it is fixed at 4 times oversampling for each pixel.
;-)
Hmm. Strange. I do belive what you say, and yet from the screenshots I have seen, 3dfx's FSAA in 2x mode looks better that NVIDIA's. Oh well, I guess 3dfx's jittered accumulation method really does give better visuals. And I thought it was just marketing...
Anyway, I don't think the V5 has much of a chance against both the GeForce 2 and the Radeon (sp?) in the marketplace. Oh well, too bad about a once good (and open-source supporting) company.
dufke
-
Yes, I know. Quake3 uses light textures, of course. But nobody would know what I was talking about if I said 'T' support... so I say T&L.
-
Somebody who knows for sure if Quake3 supports hardware T&L on the GeForce, please chime in.
Yes. I know for sure. Quake 3 uses the full OpenGL T&L pipeline. Older OpenGL games (notably Quake2) often only use OpenGL as a rasterization API, but Quake 3 uses the full pipeline. That's why Carmack was out lobbying all the hardware makers for better OpenGL driver - since if a card does not have HW T&L, the OpenGL T&L is done in the driver, which needs heavy optimizations.
-
...and the next generation of nVidia cards will do FSAA.
With the leaked drivers, this generation (GeForce) already does. It isn't as nice as 3dfx's, and comes at an equal performace hit, but it's there.
-
...call me an EE major (I am one :-) but I belive infrared is ABOVE microwave in frequency. Anyway, most of the heat produced by a CPU is transfered through heat conduction, not radiation. I don't think the actual clock signal would cause much radiation. Clocking a CPU higher may increase heat production, but the frequency of the radiation (if any) is not directly related to the clock frequency.
:-)
Anyway, even if it did, the high-amplitude microwaves in an oven are stopped by rather thin sheets of steel. And the amplitude from the CPU must be lower. Remember, your oven draws multiple kW, while I damn hope your CPU keeps below 100W. (Otherwise, you have an Alpha, and I WANT IT!
-
Would be great... but NVIDIA thought of it first, and crippled the GeForce:
:-(
It's ultra-fast with everything EXCEPT anti-aliased lines (as in wireframe models, CAD). When drawing AA lines, it resorts to software, with the result being that it is an order of magnitude slower than even a TNT.
Of course, they did it to sell more Quadra's (which is a slightly higher clocked GeForce, uncrippled, and with 64Mb DDR. *drool*).
I just thought of this... maybe this is why they don't want open drivers... the crippling might be done in software. (No, using Quadra drivers doesn't help - they most probably detect which chip is installed.)
-
Hotrail have had an SMP chipset for the Athlon available for some time
Is their chipset really done? I have heard of it many times, but never as anything else than a future product. A PC chipset isn't exactly an easy thing to make... hey, even Intel screws up at times.
-
All of the Geforce issues have been addressed (in some fashion).
Indeed. And anyway, which sane business buyer would put a GeForce in the machines...?
-
I think he mean "replaced in manufacturing" when he said "replacing".
:-)
They have to make them before thay can sell them.
-
The Spitfire *is* AMD's low end chip... it's meant to compete with the Celeries( hence the Socket-A only package )
:-).
And anyway, some tests say that even the Spitfire is faster that todays Athlon, escpecially in apps that don't demand large cache (quake3
Another reason to dump the current athlon is costs. Even if the fabs have to be tweaked to yield good on-die caches, I would say it'll still be cheaper that buying seperate cache chips from some other company.
just my 0.2 SEK
-
Dresden has been churning out 600 wafer starts per week of copper (i.e FAST!) Thunderbird all Q1..
:-)
That might seem a lot, until you consider that full capacity for Dresden is 5000 wafers per week.
(Got this from Ace's hardware)
-
The actual FPU is identical to a P2, and the x87 part is identical to P3 as well (but no SSE in Celeron 1).
The FPU might benchmark as slower because of differences in cache configuration. The Athlon has a stronger FPU than any of the Intels, but the caches are slower.
Anyway, when the poster you replied to talked about a slow FPU, he was probably comparing it to say an Alpha... *drool*
dufke
-
Bull. AGP is a direct connection to the chipset northbridge. Also: PCI is a 32-bit 33Mhz bus... AGP is a 64-bit 66Mhz, double or quad pumped if 2x or 4x. I don't know what the use of that bus would be if you had to transfer the data over PCI...
-
My GeForce 256 requires alot more wattage than the PCI bus wants to give it
:-) There are no PCI GeForces.
Well, you might want to try plugging it into the AGP bus, since that's what it's made for.
-
Slashdot will republish and resell your posts without asking for your permission.
But, in a manner of speaking, httpd does this every time someone clicks 'reload' (well, ok, not resell, but for ad profit), and I hope you agreed to that...
-