If the 9500 software mod doesn't work, how will it compare to the 9500 PRO?

[lebe0024]

Keep in mind that the 256 bit bus version of the 9500 will be faster than the 128bit version. Hopefully the mod works for me, BUT if it doesn't:

How does the 128MB 256bit Memory bus 9500 NON PRO compare to the 128MB 128bit Memory bus 9500 PRO?

There are benchies out there compairing the 64MB 128bit bus 9500 NON PRO to the 9500 PRO, but I'm sure that having the 256 bus raises the bar.

Do you think the two will perform about equal. This is relevant because some people are saying "don't risk it, just buy the 9500 PRO for $20 more", but it won't be a risk if these OEM 9500's can match the performance of the 9500 PROs.

 

[WarSong]

good question. i'm looking for the same answer.

 

[Keysplayr]

The 9500 non pro has only about 1/2 the performance at best of the 9500 pro. Tom's, and Anand have some benchmark articles.

Dont forget, you can have as much memory as you want with 128 bit or 256 bit functionality, but the 9500 non pro has only 4 pipelines to process
the data, half of the 9500 pro and 9700 (pro). You might see a small improvement for some reason or another, but I would rather have more pipes
than memory bandwidth in this case.

Keys

 

[bobcpg]

Ok, well, with AMD cpu's in the 1.7-1.8 GHz range, the 9500 NON PRO 128MB card is scoring above 10,000 3dmarks at STOCK speeds. So, I think it's quite a bit better than the regular 9500 NON PRO. I think I'm just going to start reading a lot of posts to see how well these guys perform.

 

[Killrose]

Good luck if you decide to take the plunge. But I think i'd spend the $20 extra and get the 9500 pro over the 128mb 9500 and play it safe.

 

[kmmatney]

Originally posted by: Killrose
Good luck if you decide to take the plunge. But I think i'd spend the $20 extra and get the 9500 pro over the 128mb 9500 and play it safe.

But doesn't the 9500Pro have a 128 MBit bus (so it can't be modded to a 9700 Pro)?

Here's a link showing the 9500 64 MB non-pro versus the 9500 Pro. It doesn't look pretty. However the 64 MB non-Pro version has a 128bit bus, and the 128 non-pro has a 256bit bus, so the 128 non-pro might not be so bad.

9500-9700

Also, as bobcpg said above, people's 9500 non-pro 128MB 3dMark scores at look pretty good even at stock speeds.

 

[D3xx]

The Xbit labs review seems to indicate that the 9500 pro is 50% faster than the
9500 np and that the 9700 np is 70% faster than the 9500 np. However those
who have successfully modded their 9500np and posted benchmarks indicate
that the (effectively) 9700np is only 30% faster than the 9500np.
So I'd expect a 9500 pro to be 20% faster than a np in reality.

 

[lebe0024]

Ok, using posts from another thread in this video forum, overclockers.com forums, and the Online Result Browser on Madonion.com, I was able to find that for a stock Athlon XP @ ~ 1.8 - 1.85 GHZ you will expect to get about:

~10,600 for a STOCK 128MB 9500 Non Pro with 256bit memory bus.
I got this number simply from people's before/after postings about the mod. Some included CPU speed, and surprizingly, many were using an athlon XP at this speed. So, I took an average of these.

~10,900 for a STOCK 128MB 9500 PRO
With the orb, it's hard to tell who's using what card in the 9500/9700 series, because the ORB itself doesn't differentiate. You have to go with only the people who actually put there card and it's speeds in the project description. So, I found all of them for this processor, and took an average. There was an instance where a guy got 11,417 with his stock 9500 Pro, but his FSB was at 200. Most normal FSB speeds (133 - 166)were around 10,900.

My findings could be wrong. But that only looks like a 300 point difference. I encourage others to do the same. Look at people's before and after postings with CPU speed, and check the orb for Stock 9500 Pro scores at those same cpu speeds.

 

[jaybee]

This may be a touchy subject, but geez I am tired of seeing 3dmark results at 10x7 with no AA. Not many people are going to run a 9500 Pro or better with those settings. I don't run anything without AA anymore. btw I am using a 9500np for the last week or two, GF4 4400 before that. Anyway, comparing the 9500Pro to a 9700 model at 10x7xnothing is a bit silly IMO. Run the benches at settings you would use with the cards. 10x7x4AAx8AF seems more appropriate to me.

jaybee

 

[WarSong]

Okay, i took the plunge. I'm a gambling man so I orders up the 9500np from Newegg. I'll let you guys know how it goes. Wish me luck

Originally posted by: jaybee
This may be a touchy subject, but geez I am tired of seeing 3dmark results at 10x7 with no AA. Not many people are going to run a 9500 Pro or better with those settings. I don't run anything without AA anymore. btw I am using a 9500np for the last week or two, GF4 4400 before that. Anyway, comparing the 9500Pro to a 9700 model at 10x7xnothing is a bit silly IMO. Run the benches at settings you would use with the cards. 10x7x4AAx8AF seems more appropriate to me.

jaybee

They run the benchmark at the default settings so they are comparible to other peoples results. If everyone used different settings then the benchmark would be useless. apples to apples.

 

[Killrose]

Good luck Warsong!! Bring those numbers up on the success side of the 9500 to a 9700 poll OK?

 

[lebe0024]

This may be a touchy subject, but geez I am tired of seeing 3dmark results at 10x7 with no AA. Not many people are going to run a 9500 Pro or better with those settings. I don't run anything without AA anymore. btw I am using a 9500np for the last week or two, GF4 4400 before that. Anyway, comparing the 9500Pro to a 9700 model at 10x7xnothing is a bit silly IMO. Run the benches at settings you would use with the cards. 10x7x4AAx8AF seems more appropriate to me.

Dude, read the first post. If I had the cards to benchmark, I would do it. If you have both cards, let us know. But, seeing as there are no reviews for the 128MB 9500, and especially no reviews compairing it to the 9500 Pro, we're trying to do our best to piece together a clear picture of performance.

What kind of 9500 np do you have? And how did it compare to your ti4400 ? That would be a great help.

If any of you have this 128MB 9500 non pro UN MODDED and UN-OVERCLOCKED, run us some benchmarks at typical CPU speeds for some AA and AF in games and 3dMark.

 

[jaybee]

Glad to help. Go to the 3dmark Search & Compare and select jdbuhr@bellsouth.net as the user email. You'll find tests at different settings for the 4400 and 9500np. The 4400 scores quite a bit higher on my system at the settings used.

jaybee

PS I ran the 9500np at settings I expected I would use: 16x12x0AA, 12x9x2AA, and 8x6x4AAx8AF. I get some decent improvements by overclocking, but still can't touch the 4400.

 

[lebe0024]

Jaybee,

I did the search and compare, and only 4400 scores showed up (and one 8500 scores). There were no 9500 scores listed. Also, what kind of 9500 do you have? How many megs of ram?

 

[Mingon]

how about 12k 3dmarks with core @350

 

[WarSong]

Originally posted by: Mingon
how about 12k 3dmarks with core @350

What about the rest of your system specs?

 

[Mingon]

have a look

 

[lebe0024]

I think that if you're mod is unsuccessful, you gain the most performance out of overclocking your core. (which makes sense, because it's the core that's the bottleneck.)

Jaybee, I found your scores, but it was hard to find a set of benchmarks for 4400 and 9500 where the same settings were used. And ALL of you scores seem surprizingly low.

 

[jaybee]

Originally posted by: jaybee
Glad to help. Go to the 3dmark Search & Compare and select jdbuhr@bellsouth.net as the user email. You'll find tests at different settings for the 4400 and 9500np. The 4400 scores quite a bit higher on my system at the settings used.

jaybee

PS I ran the 9500np at settings I expected I would use: 16x12x0AA, 12x9x2AA, and 8x6x4AAx8AF. I get some decent improvements by overclocking, but still can't touch the 4400.

System details - Athlon XP overclocked to 1600-1680MHz. 512MB PC133. OS: XP.

 

[WarSong]

Check out this: Link