Sign in with
Sign up | Sign in
Your question

Celeron Upgrade Help pls!

Tags:
  • CPUs
  • Celeron
  • Pentium
Last response: in CPUs
Share
September 22, 2001 12:27:47 AM

Hi, I need alittle info on the major differences between the celeron processor and the pentium? Basically I was looking to squeeze out some more use from my current PC before I go ahead and buy a new one at the end of this year and pass this one over to my boy. Will upgrading to say a 1 ghz celeron have that much of a difference vs the 450 pentium I have now?

Current system specs:
Asus P2B-F mobo
Pentium III 450 mhz cpu
512 MB SDram 100 (upgrade from 128 since it was so cheap)
geforce 2 MX 200 (upgrade from a diamond viper 770)

More about : celeron upgrade pls

a b à CPUs
September 22, 2001 12:56:26 AM

It would be a little faster. If you live in the U.S., the PIII 700 goes for about $6 less and performs on par with a Celron 950. But here's the kicker-the PIII 700 can be easily overclocked to 933, where it will beet ANY speed Celeron.

Back to you Tom...
a b à CPUs
September 22, 2001 12:58:38 AM

Oh, is your MX200 faster than your Viper 770? I would really be surprised considering the MX200 has a 64-bit SDR memory interface, but the Viper 770 has a 128-bit SDR interface (and the 770 Ultra had a 128-bit DDR interface!).

Back to you Tom...
Related resources
September 22, 2001 1:45:41 AM

I did notice that after you pointed it out to me but the card specs for the MX were 256 bit graphics architecture, 32 MB DDR ram which I figured was an upgrade from the viper 770 (not the ultra)...
a b à CPUs
September 22, 2001 3:49:21 AM

Well, if your 770 was only 16MB you might see a very slight improvement.

Back to you Tom...
September 22, 2001 11:30:17 AM

The gf2 is much faster than a v770, I used to have a v770 ultra, it generally scores 2000 or so in 3dmark 2000(the older one) gf2's (even mx's) generally score in the 4000 range. Even though 3dmark should never be used alone, I think those scores mean something.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
a b à CPUs
September 22, 2001 11:59:40 AM

I just got curious so I put a V770 (not Ultra) in a system previously tested with an MX200 and got a better score with the TNT2 card. In fact, I got about the same score with a TNT2/Celeron 850 as I did with an MX200/PIII 933! Memory is not perfect, so maybe someone will look up my old post and find out what my MX200 score was. I think my scores were as follows:
MX (original, like the MX400), on PIII 933: ~5800
Same card on a Celeron 850................: ~4500

MX 200, on a PIII 933.....................: ~3200
Same card on a Celeron 850................: ~2800

Now heres the kicker:
TNT2 (via Viper 770), Celeron 850.........: 3285

I was VERY IMPRESSED with the performance of the V770, considering it's age. It does have a 128-bit memory path going to 6ns Ram. All these scores are for overclocked cards at their maximum stable speeds, which were, to the best of my knowledge, 220/190 for the MX200, 220/190 for the MX, and 150/185 for the TNT2 (older memory modules retain heat better, 190 only worked for a few minutes but 185 was stable).
Oh, and there was NO JERKINESS in the later parts of the test (the ones with the rope coil looked especially nice, as did the one in the tube). I always get some jerkiness with my newer cards, even a stock clocks. The Helicopter tests were the ones were the GeForce cards did the best over the TNT2.


Back to you Tom...
September 22, 2001 1:55:22 PM

Interesting, do you remember what driver revisions you were using, because madonions score average seems to dissagree with your assessment.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
September 22, 2001 2:03:44 PM

OK I looked it up on madonion, the ULTRA highest score for a v770 Ultra (tnt2 ultra) was 4200

the HIGHEST for a gf2 mx was 8011

these were all on p3 systems, I did not search for equal processors however, hold on.


a p3950 gf2 mx got 6772(highest)
and the average of the first page is 6000

a p3 950 with a v770 ultra got4215(higest)
and the average of the first page was 3800

The results are waiting at madonion, the gf2 mx is a better videocard than the v770.




~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
a b à CPUs
September 22, 2001 8:51:45 PM

You don't seem to understand that the MX and MX400 are about 50-100% faster than the MX200, and 3D-Mark does NOT differentiate! They don't differentiate between the GTS and PRO either, which puts the PRO at an unfair advantage since it has faster memory. I'd ask you to look up the scores on a MX200, but you CAN'T, their mixed in with all the MX and MX400 score with no way to know which is which! Also many of those MX's on the first page are the new super fast GAinward cards with 3.5ns memory! The MX200 uses 6ns memory, so it not only has half the bus width, but at half the memory clock of those cards!
Oh, the driver revision is 6.31, which works very well in spite of what you might think about it's age (I couldn't get more than about 5% better with the newest Det4 driver, and that driver causes high-speed jerkyness).

Back to you Tom...
September 23, 2001 1:13:18 AM

Even so, the mx in general is a better card than the v770 I checked the ULTRA scores, and they were much less, I did not intend to discredit you or anything, I was just showing the 3dmark scores I saw myself. It may be true that 3dmark does not seperate the mx 200 from the 400, however the averages speak for themselves.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
a b à CPUs
September 23, 2001 1:59:03 AM

NO, averages don't speak for themselve, take a class in statistics. For example, if braindead people have an IQ of 0, and I had a brainded person in my family, I was IQ 140, and the other three members of my family dead average at 100, the avaerage IQ for my family would be 88. If you tried to apply that average to me it would give you a very inacurate view. Most people who post at MadOnion are performance freaks. How many of them do you think own the MX200? How many of them an "ultra fast" version of the MX (such as gainwards)? Would it be safe to say such a performnace freak would be more likely to buy the fastest card?

Back to you Tom...
a b à CPUs
September 23, 2001 2:07:52 AM

Oh, and a Corvette is generally better than a Cavalier, but if it's got 5 dead pistons the Cavalier will win the race! The MX200 is junk!

Back to you Tom...
September 23, 2001 4:33:49 AM

Crashman, that same law applies to the tnt2 scores on madonion! So that argument applies both ways, I got both sets of scores from madonion. I agree that the mx200 is not differentiated at madonion, which makes those results not entierly accurate, as I said my post was not meant to say you were wrong, I basically was asking you why madonions scores differ from your reported scores. Which you explained, I still however believe that the mx200 is a better card than the tnt2 (not ultra) if even only barely.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
a b à CPUs
September 23, 2001 5:37:01 AM

I know for sure that the MX200 lost on MY spare system in THAT benchmark. If I still had the MX200 I would run some other graphics benchmarks and post a review in the Graphics Cards forum. But since I don't, I can only go by the one test I did run, 3d Mark 2000.
You do have to admit that a 128-bit memory interface is a HUGE advantage for the TNT2, even if you believe that it can't offset the weaker graphics processor.

Back to you Tom...
September 23, 2001 9:28:00 AM

I will agree to that, it is double the performance nearly. I am now intrigued and will look up more benchmarks online to find out which card is better, if a gf2 is outperformed by a tnt2 that would make it the red headed bitch of videocards LoL.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
Anonymous
a b à CPUs
September 23, 2001 3:30:49 PM

this thread was about celerons, right? :) 
a b à CPUs
September 23, 2001 7:30:50 PM

Is the TNT2 (NOT the M64 with similar castration, but the regular version) still in production? Because if it is, and it beets the MX200, this information could be very bad news for MX200 sales, should it become common knowlege!

Back to you Tom...
September 23, 2001 8:32:18 PM

Of course the MX200 is faster then the tnt2 ultra
considering the tnt2 ultra was a bit slower then the original geforce 256, and on par with the voodoo 3 3000
I have owned all 3
My Guillemot Maxigamer Xentor 32 was one of the fastest ultras aside from the Hercules version and the Asus 256 beat me switching both with same Detonators, and my wifes MX regular smoked my geforce 256 and tnt2 ultra

Blame the newbies not the technology
a b à CPUs
September 23, 2001 8:50:56 PM

Read my above post, the REGULAR TNT2 (Ultra not needed!) BEETS the MX200 in 3D-Mark 2000! And the Ultra is even better, the MX200 sucks!

Back to you Tom...
September 23, 2001 9:29:47 PM

If an MX200 can beat an ultra , it's not likely that a regular tnt2 can beat an mx200

Blame the newbies not the technology
September 23, 2001 10:26:02 PM

just curious: you said "your wife's MX regular". Is that an original MX (=MX400) or indeed an MX200?
September 23, 2001 10:38:01 PM

She had a regular MX and the other machines I benched had the MX200 32mg and 64mg both beating the Diamond Viper770 ultra and maxigamer xentor 32 Me and my boss are gamers and screw around with HW when there are no machines to be done..
Granted the MX200 is crippled, but is indeed faster than the TNT family of cards,
sorry for not being as clear as I could have been

Blame the newbies not the technology
a b à CPUs
September 23, 2001 11:23:58 PM

An MX200 CANNOT beet a NORMAL TNT2, let alone an Ultra. I have the test scores myself ON THE SAME DRIVER!!!!!! WTF was wrong with your TNT2 I'll never know, but I got signifiaclntly better scores with the TNT2. I also tried a regular MX which stomped both the MX200 and the TNT2. Again on the same system/same driver.

Back to you Tom...
September 23, 2001 11:31:32 PM

Settle down there fella... I am only talking here, grow up..
I am only stating my opinion here and that is what I found.
So what if I dont agree with you. And yes I think an MX200 will beat a tnt2

Blame the newbies not the technology
a b à CPUs
September 23, 2001 11:41:03 PM

I don't need to "grow up" anymore. I need you to quit confusing people with your opinions when I present them with facts. I did benchmarks, simple factual information. For you to interject a conflicting opinion serves no purpose other than to perhaps boost sales of an inferior product.

Back to you Tom...
September 23, 2001 11:49:42 PM

I have sold the MX200 64mb to many gamers as a value gamers card and all of my clients really enjoy the performance...
And yes when I was testing the cards I did not make any HARD evidence... But as you should already know every dicussion has oppositions, I am not trying to insult you I am just voicing an opinion

Blame the newbies not the technology
September 24, 2001 1:45:08 AM

Chill crash, I have some questions on your test, which resolution did you run the test at, color depth?

I personally think the gf2 mx200 would beat the tnt2, I had a tnt2 ultra and I could barely scratch 3000 in 3dmark2000, mx200's apparently are hitting 4000-5000. I dont doubt you did tests but I wonder what would lead your tests to differ from others so greatly.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
a b à CPUs
September 24, 2001 4:41:16 AM

1024x768x16, yes I always run the default test for comparisons. OK, I haven't seen anyone specifically post their 3D-Mark 2000 score for the <font color=red>MX200</font color=red>. All I've seen is MadOnions database which do not differentiate the fastest (3.5ns MX400) from the slowest (MX200) cards. If anyone want's to post an MX200 score specifically, go ahead! You have proven nothing, except that the FASTEST MX400 cards are giving great score, which does not imploy in any way that the MX200 can be compared. And this other guy is probably only arguing to boost his MX200 sales!

Back to you Tom...
September 24, 2001 6:16:05 AM

Like I said, crash, calm down, your post tone is comning off as angry.

The 3dmark2000 database I admit does not differentiate, I am just trying to get to the bottom of the situation.

Thanks for posting the resolution, do you have the driver revision you used. Thanks.

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
a b à CPUs
September 24, 2001 7:21:25 AM

Look 22 post up from this one, it's 6.31 for all cards, it's the same system.

The MX400 gives about 4500 with a Celeron 850

The MX200 gives around 2800 with a Celeron 850

The TNT2 gives around 3200 with a Celeron 850

The same system gave a 5800 with the MX400 on a PIII 933, which is about right, is it not? Oh, and averaged 7200 for a GTS and a 933.

And all this time I thought the TNT2 didn't work well with det3 drivers. Must've been resolved in this revision.

And the final thing I mentioned, should you need a reminder, is that ALL CARDS were overclocked to their best performing speed, which was around 220/190 for the MX's and 150/185 for the TNT2. So the MX even had it's memory clocked higher!



Back to you Tom...
September 24, 2001 9:03:09 AM

Ok, I believe you, one final question.

How much of a % increase from stock were those overclocks, and what kind of core overclocking did you do on both.

If those results are not flukes(and I think they arent, crash has always been a good source of info) then the mx200 is the red headed bitch of all videocards.

::crowns the mx200:: all hail red headed bitch of video cards. ;-)

Does anyone else have a mx200 they can post 3dmark2000 scores from, for some averaging.



~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~<P ID="edit"><FONT SIZE=-1><EM>Edited by Matisaro on 09/24/01 02:07 AM.</EM></FONT></P>
September 24, 2001 2:50:21 PM

"this other guy is probably only arguing to boost his MX200 sales!"
I only sell this card to people who cant afford a GTS or PRO
on a Duron system. I am not trying to boost any sales of any inferior product... My store buys these cards for $60 canadian, and for that price the performance is pretty good.
When I was benching I did all tests at stock speed no O/C.
My ultra was stock 175/190 and got spanked by a 64mb mx200



Blame the newbies not the technology
a b à CPUs
September 24, 2001 6:31:12 PM

I ran all the test on both processors for the MX and MX200 (and the regular MX did fairly good!). As for the TNT2, I only ran that one on the Celly, because I didn't feel like pulling my processor for it. No need to do that anyway, the TNT2 did good enough on the Celly to pass the MX200 on either.
Yes, I would like to see some MX200 specific 3D-Mark 2000 scores besides my own.
Oh, the stock clock for the TNT2 was 125/150 I believe. I don't pay too much attention to stock clocks anymore, the fact remains that all the cards were overclocked to their highest stable performance levels.

Back to you Tom...
a b à CPUs
September 24, 2001 6:39:21 PM

Yep, thanks for the link! The TNT2 has a 2.4GB/s memory bandwidth, the MX200 has only 1.3GB/s. Since most 3D apps are limmited mostly by memory bandwidth, your helping me proove my point!
All of us with GeForce 2 cards know that these peak pixels per second rates are meaningless if your memory cannot supply and adequate data flow to the GPU. Otherwise, the MX200 and MX400 would score the same! The main difference between the GTS, Pro, and Ultra is the memory speed. We all know that increasing your memory speed with these cards by a certain percent will boost your scores by a similar percent on these cards. So cutting your memory bus width in half will cut your scores nearly in half. But since the memory bnadwidth is already a significant limmitng factor on the MX400, the MX200 further compounds these problems. Evidence of this can be seen in the 3.5ns Gainward MX card, which has superior scores over even the MX400 on which it's based. The MX (even the Gainward 3.5ns model) will NEVER achieve it's peak rate because of these limmitations. Which literally means that peak rates get thrown out, memory speed is the ONLY factor in speed on these cards.

Back to you Tom...
September 24, 2001 7:04:16 PM

Also look at the pixels per second....
If you are going by memory bandwidth than the MX400 would lose to the tnt2 ultra...
Memory bandwidth as you should already know is only one factor when benchiing performance numbers...
:: TNT2
Memory Speed: 150 MHz
Pixels Per Second: 250 Million
Memory Bandwidth: 2.4GB/s

:: GeForce2 MX 200
Memory Interface: 64-bit SDR
Texels per Second: 700 Million
Memory Bandwidth: 1.3GB/s

Look at the texels per second


Blame the newbies not the technology
September 25, 2001 6:12:20 AM

Memory bandwidth is not the only limiting factor, and it dosent have much weight when it is on different platforms. (voodoo3's had alot of mem bandwidth too remember).

~Matisaro~
"The Cash Left In My Pocket,The BEST Benchmark"
~Tbird1.3@1.5~
!