• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

My old 939 vs AM2 = Closer than Im happy with

FatRakoon

FatRakoon

FatRakoon

Soldato
Joined
18 Oct 2002
Posts
10,475
Location
Behind you... Naked!
Until only a few months ago, my fastest AMD was my Opteron 175

Asus A8N-SLI & Opteron 175
2 x 7800GTX
Creative XFI
4 x 1GB Corsair XMS 3200 TwinX

Now, my main Intel is a Q9550 and my No2 PC was an Intel Q6600. Obviously both quads, and so, I set about swapping my No2 Intel for a half decent AMD.

I was offerent a Phenom 9550 and a DS3 Mobo, and since I love the DS3 Mobo, and I have 3 of them, and in fact the Q6600 was also in a DS3 too, I thought why not... a DS3 + Quad core Intel for a DS3 + Quad Core AMD

I knew I would be dropping in speed, that is not the issue. I dont care too much for the speeds because the PCs are only used as mess abouts and on the LAN anyway... I have my main PC for when I want speed ( Never have done to date to be honest )

Now, the things is this....

I know I am using a benchmark thats probably out of date, but when I run SuperPI MOD on my Opteron 175, I am getting 39seconds on the 1M run. Now, on the Phenom Quad Core, I got 36 Seconds. WHAAT?

Ok, the Opteron is at 2.0 and the Phenom at 2.2 and so I clocked the Opteron to 2.2, the same as the Phenom and what do I get? 35 Seconds???

So, clock for clock, my Socket 939 Opteron running DDR memory, is actually faster than my phenom thats running DDR2?

How is this possible?

So, I quickly run the 1M test on my Intel and I slap 19 Seconds, so thats fine.

These are all NOT clocked... Motherboard and RAM are tweaked yes, but no actual overclocking at all ( Except to do that one run with the Opteron at 2.2 )

Now, the things is that I am looking at my semi-spare lists and I have a spare 775 Mobo and a spare TriCore AMD AM2 and its clear that spending on a Mobo for the AMD is a total waste of time, and instead I should simply buy the most naffest slowest Intel... And it will be quicker than this extra Tri Core AMD wont it?

What a shame.
 
SuperPi is hugely Intel biased.

You can get weedy little Intel chips outdoing the best AMD has on that benchmark.

That doesn't mean the weedy Intel chip is actually better.
 
Not even sure the Super Pi is multi - threaded, if its not, you're throwing away multiple cores worth of power.
Do yourself a favour & run Cinebench 11.5 or some other real benchmarks to see real performance.
 
No, this is true.

My real point of this is that the old 939 Opteron is in fact faster than my AM2 Phenom Clock for clock

Can oyu suggest a truly fair benchmark that I can run on old and new, intel and AMD?

Does such a thing truly exist because every benchmark apps that I have had always leans one way or another... Perhaps maybe Performance Test is about the most neutral maybe?
 
Quixote said:
Not even sure the Super Pi is multi - threaded, if its not, you're throwing away multiple cores worth of power.
Do yourself a favour & run Cinebench 11.5 or some other real benchmarks to see real performance.
Click to expand...

Sorry I didnt see that one.

No, SuperPI is not...

If it was and my Dual core Opteron was still beating the Quad Core Phenom then Id really be piddled off LOL, but no, its not.

Again some other REAL benchmarks?

Ok, CineBench. I will try that...


---

Ok, I just run CineBench

1 - I need to install Video Drivers ( Shame because I am also playing Quake 2 and that dont like the newer drivers )

2 - I run the CPU test and compared to teh other 5 examples, I came not only last, but embarrasingly last too!

Never mind I will compare my own systems with it and see how bad they fare against each other.

Thanks for the info on that one... Any others you can suggest?
 
Last edited:
Ok, I just run CineTest on my main PCs now... Bar one.

The Gap between the AMD and the intels are indeed closer, and the Opteron has shown itself to be around half the speed of the Phenom which is what I had originally hoped it would be before running Super PI

One annoying thing however is that I have a few 8800 based cards and one 250GTX and this benchtest does nothing to help them out.

I have

Point Of View 8800 GTS 320MB ( 2 of, but Not in same machine )
Point Of View 8800 GTS 640MB
ACER 8800 GT v512MB
Palit 250 GTX 1GB

And the slowest is the 8800GTS coming in at 23.8 on the Open GL Score, when the fastest is the 250GTX coming in at a rather irritating 27.6

This is ONLY proving to me that upgrtading from an 8800 to a 250GTX is near a complete waste of time.

I had to reinstall all the nVidia drivers so thats piddled me off, and I have not tested the 280 nor have I tested any ATI Cards either. I will do that lot in teh next few days.
 
Well... on that score, a 250 isn't a great upgrade from an 8800. Isn't it just a re-launch of the 9800GTX+? Given that the 9800GTX+ itself is a revision (albeit shrunk and clocked) of the same G92 core as in the 8800GTS, you wouldn't expect to see a big boost replacing one with the other really, I wouldn't have thought. The clock boost and extra memory will help, of course, but that probably accounts for the fairly small jump you're seeing.
 
I get 20 seconds on super PI using 1 meg test. Im gonna try cinebench.. im running a Intel Q9400 @ stock at the mo. Ram is 6400 running @ 500mhz though.
 
Last edited:
Hmm that annoying my score was - 2.89cpu and 25.10 fps on the opengl. not good? i think i might need to change a few settings in my catalyst settings and try again, i think ive got vsynch forced on? anyway whats up wi that score??
 
phenoms dont really like superPI for some reason, my old single core opteron wasnt my slower than my phenom, but then my phenom can run 4 threads at once. running at threads it's quicker still...


..but then there's other benchmarks where the phenom is a LOT quicker so meh. i dont really care :p
 
I have always known that benchmarks dont truly give you a true comparison.

One banchmark that I did like for a while was Aquamark3 and even today it still gives a fairly respectable answer thats a lot closer to the truth than some other banchmarks,

For example, I remember getting around 25K on oen card and 48K on another and the newer card was indeed about twice the speed in all games and my average FPS went from around 40 to around 75-80 so that score was prtty much spot on.

The exact same thing happened when I got a score of 70K and 80K, the percentages were absolutely spot on in the games i was playing too, while 3DMark was massively way off.

Now, benchmarks I feel are only actually any true cop when you have a PC and you run the test at stock, and then you tweak your PC and keep re-running the test to see how much that PC has improved, but other than that....you really cannot possibly use only one benchmark to get a fair comparison can you?
 
You must log in or register to reply here.
Back
Top Bottom