Printable View
1ghz RAM??????Quote:
Originally posted by dimava
thats not the best computer.... a Server with four Pentium 4 1.5Ghz chips and 200mb (ok kb, i froget) cache, and 1ghz ram
the one above isn't the best, but its pretty good
Athlon is faster than the PIV(BTW, IV is the Roman Numeral for 4). The PIV has more MHZ, but the Athlon does more stuff per MHZ...Quote:
P IIII is faster then Athlon
P III is slower then athlon
You'll probably get banned from the board if you don't change your signature :)
I don't think John likes stuff like that.
Anyway, if you get technical, pentium is latin for 5.
If the p-iv is really a new architecture, then naming it Pentium ANYTHING is just plain silly.
I think the p-IV does more stuff per mhz, but the problem is that most of the stuff it does is useless. It tries to anticipate what it thinks you want it to do before you ask it, and pre-calculates values into a buffer. This is one of the reasons it is slower than a p-III of the same speed. It wastes processing power.
And FYI, the latest Athalons are multiple-processor compliant. I have yet to see a desktop motherboard that uses the ability though. When i DO, ill be sure and snatch that sucka up. Hell, even the 700mhz mac g4 eats p-IV for breakfast. Imagine what 2 1.2 gig athalons overclocked to about 1.4 will do. ;) Seriously, i'm really waiting for the server chip to come out with the 2 meg of cache ram :)
Ok. I'm missing something here. You say that there is almost no difference with a PIII 933 and a T-Bird 1000. I'll agree to that with your configuration. The thing with intel CPU's is that they work much better when put onto an intel board (not just intel chipset). You'll see a drastic improvement. The only problem with the intel boards is that you can't overclock at all.
heh heh can't overclock on intel boards... That is just so wrong. Buy a slot-1 board and put a slocket adapter in it. Buy a 800 mhz chip designed for 100mhz fsb, then set the slocket jumper to force operation at 133 mhz. Viola, you have a 1066 mhz. If you don't think this is even possible, consider that i have overclocked EVERY pentium i have ever had (and every athalon and t-bird except one), even the Pro on the IBM board.
If the board has jumpers you can change them. If it doesn't, you give it jumpers.
Also, i gave wrong info on the pentium system. The ibm chipset doesn't support agp 4x or a multiplier divider for a 133mhz fsb. So my AGP was actually 2x but running at about 42 mhz instead of 33. Don't try this with a pansy video card.
The 2 Configurations:
Intel boards have no jumpers and no settings in BIOS either. It auto-detects everything on startup.Code:Intel AMD
m/b Intel D815EEA Gigabyte GA-7ZM
Ram 256 (133mhz) 256 (133mhz)
Hdd 2x 30gig Fujitsu 1x 30gig Fujitsu
V/C Diamond Stealth Nvidia Geforce
S540 Extreme
S/C SB Live! Player SB Live! Value
AGP 4x 4x (???)
I know its possible to overclock. The reason I don't do it is because I am given a new CPU every 4-5 months and must then return the old CPU to suppliers who then re-package it and resell it. So I can't exactly shorten the life span!
well you didn't mention the processor speeds, but just the fact that the video card is different and you get different results in 3d games tells me that it is probably the 3d card that gives you the better performance.
The gigabyte is a :ahem: decent motherboard. I never have understood why it needed a one meg bios though. Does it have a startup graphic?
Seriously, and no bs involved, you should retest with the same vid card in both. I think you will be very surprised of the result.
Processor speeds was mentioned above. PIII is a 650 the T-Bird is a 700.
Do you not read my posts? Let me explain the procedure.
First. Test PIII with Stealth. Then, switch off. Remove stealth. Plug Geforce in. Retest
Then, test T-Bird with Geforce. Switch off. Remove Geforce. Plug Stealth in. Retest.
We compare results with same v/c and PIII wins. Not by much though. Still, we are comparing a 650 to a 700.
Slight correction: I looked over my old notes and saw that i gave some slight errors in my benchmarks. I stated that i was getting results in the 5300s on both machines. This was correct. What was wrong was i said it dropped to 5200 with the duron. I found my saved info and to verify it i retested my system today. It actually dropped to 4900.
However i have since this afternoon trounced every reading. The geforce2 is really an overclocked geforce256, so i thought "i'll overclock my geforce256"
And that is just what i did. I bumped my processor speed up from 750 to 800, then raised the geforce processor speed to 260 from 200 and the geforce memory speed from 300 to 330 (i could have went farther up to 390) and retested in 3dmark. Now i am getting in the 5700s.
Sounds like your benchmark is based mainly on your graphics board. You should run tests with direct draw rather than D3D or OpenGL.
The Geforce 2 MX is an overclocked Geforce 256. The GTS has 2 extra pipelines if I can remember correctly. It has a different GPU as well.
you can do that with 3dmark. Just change the setting of optimizations to "direct3d software t and l"
I no longer use 3dmark2000 :) 3dmark 2001 just came out! :D
I have a thread about it in the chitchat area :)
to download it, go to www.madonion.com
The demo on both programs is amazing.
Anyway, i also mentioned earlier that i also did benchmarking with norton utilities, which ignores video cards. For the best benchmarking available, however, check out sisoft sandra2000. It has a comparison chart as well.
I didnt post the results of that test since i no longer have the program and can't remember them.
And doesn't Unreal Tournament use direct3d? I know i have Unreal set to Direct3d on my system. OpenGL is probably even more Video card specific than direct3d on performance hits though. Since the specification states that an open gl driver has to emulate in software functions that the card doesn't have.
Yep, but so does Direct3D :)
ok try this one a server with 32!! 733pentium Xeons 32Gb RAM!! and 2-3Tbs!!!! each proccessor having 2Mb cacheQuote:
Originally posted by dimava
thats not the best computer.... a Server with four Pentium 4 1.5Ghz chips and 200mb (ok kb, i froget) cache, and 1ghz ram
the one above isn't the best, but its pretty good
Okay...what about an SGI workstation with 512 processors?
Quote:
Originally posted by parksie
Okay...what about an SGI workstation with 512 processors?
:eek: *Dennis faints*
A video card's Direct 3d driver does NOT have to include functions that the card doesn't support. It just tells directx that the function is unsupported. Example: Directx says: Scale and rotate this image.
Card driver says: Card doesn't support scaling.
OpenGL says: Scale and rotate the image.
Card driver (determines that card doesn't support function, then emulates it with software): Ok.
That is what is so nice about opengl. It is possible to run 3d applications on it on ANY video card. Even ones without 3d accelerators. Your processor will take up the slack however. For an example, see the open gl screen savers, which work on any video card, even piece-of-sh** s-3 mach 64 cards, which don't support 3d. Actually this is a better card than the ati Rage 2 agp cards, which REALLY bite.
Unless there is something inside of DirectX 8 that i don't know about, this is the reason more and more developers are writing their code for open-gl. You don't have to worry about compatibility.
Yea......ok........ :rolleyes:Quote:
Originally posted by parksie
Okay...what about an SGI workstation with 512 processors?
D3D will, in many cases, have software support (the Hardware Emulation Layer) that does it, but it's extremely slow (a lot slower than OpenGL) and low quality. Also, OpenGL is capable of mixing hardware/software together :cool: You may have said that but just in case I misinterpreted :rolleyes:Quote:
Originally posted by Lord Orwell
A video card's Direct 3d driver does NOT have to include functions that the card doesn't support. It just tells directx that the function is unsupported. Example: Directx says: Scale and rotate this image.
Card driver says: Card doesn't support scaling.
Either way...OpenGL is a LOT easier to program for...in agreement with whoever said it earlier :D
UT runs anything from DirectDraw to Glide. OpenGL is also video card specific, that is why I said you should run a directdraw test instead.Quote:
Originally posted by Lord Orwell
And doesn't Unreal Tournament use direct3d? I know i have Unreal set to Direct3d on my system. OpenGL is probably even more Video card specific than direct3d on performance hits though. Since the specification states that an open gl driver has to emulate in software functions that the card doesn't have.
The opengl driver that comes with windows works with any card. It just doesn't work as well as a card-optimized one.
I have a video card that directdraw fails on.
I have two that direct3d fails on.
open gl works on all of them.
here is an excerpt from the readme file for a direct3d game:
Glitch Disable
------------------------------------------------------------
-green boxes surrounding Software Transparent BLT
any object in the game
-box surrounding mouse cursor
-animations appearing Software Mirror BLT
'reversed'
-Fog-of-War appears in
a 'semi-checkerboard'
pattern
-fog-of-War has edges
that look 'spiky'
-vertical lines down Software Standard BLT
the main game screen
-the game Cursor is
leaving 'artifacts' of
itself on the game screen
-any other non-resolved
graphics problem
Specific Video Cards and Fixes
------------------------------------------------------------
Matrox Millenium (1) : Enable Software Mirror and Software
Transparent BLTs.
Permidia : Enable Software Standard BLT.
Fair enough