My own complete review of Phenom 9500

In another thread (author was banned, for multi-nick login) some peoples want my results ... here is:

My complete review in Czech language: http://www.pctuning.cz/index.php?opt...9815&Itemid=28

Some results photos, and config ...

http://www.pctuning.cz/ilustrace3/ob...0/Phenom_8.jpg
http://www.pctuning.cz/ilustrace3/ob...0/Phenom_6.jpg
http://www.pctuning.cz/ilustrace3/ob.../Phenom_1b.jpg
http://www.pctuning.cz/ilustrace3/ob...500/Cpuz_1.png

Config and testing systems ...

http://www.pctuning.cz/ilustrace3/ob...est_system.jpg
http://www.pctuning.cz/ilustrace3/ob...est_system.jpg

http://i13.tinypic.com/8dzokte.png

OC ...

Only 230MHz HT frequency, BIOS and mobo is flaky ... i had never seen worst mobo ever ...

http://www.pctuning.cz/ilustrace3/ob...500/Cpuz_4.png

This mobo with last F2L BIOS dont have a CPU Multiplier change in BIOS! It it strange ...

Consumption ...

First is Load, second Idle ...

http://www.pctuning.cz/ilustrace3/ob...0/spotreba.png
http://www.pctuning.cz/ilustrace3/ob...spotreba_2.png

Tests ...

http://www.pctuning.cz/ilustrace3/ob...inebench_2.png
http://www.pctuning.cz/ilustrace3/ob...500/winrar.png
http://www.pctuning.cz/ilustrace3/ob...0/3dmark06.png
http://www.pctuning.cz/ilustrace3/ob...crysis_cpu.png
http://www.pctuning.cz/ilustrace3/ob...erest_read.png
http://www.pctuning.cz/ilustrace3/ob...crysis_800.png
http://www.pctuning.cz/ilustrace3/ob...0/bioshock.png
etc ...

Look at complete review, more tests and games ...http://www.pctuning.cz/index.php?opt...9815&Itemid=28

Can Phenom to compete to Intel? ... NO! yet ...

Intel has : Great Overclocking, better performance, lower consumption ...
Phenom has : No Overclocking, lower performance (at the same clock), brutal consumption ...

PHENOM is C**P! ... till new revision, B3 - B4?

ORB -- I need to say this....

One heck of a fine (outstanding) job.

Quote:

---

Only 230MHz HT frequency...

BIOS dont have a CPU Multiplier change in BIOS! ...

---

same prob with you , can't change multiplier in bios :confused: ,

can't run AMD Overdrive :confused:

Quote:

---

Originally Posted by iMAX

same prob with you , can't change multiplier in bios :confused: ,

can't run AMD Overdrive :confused:

---

I think you are confused ... the multipliers on this CPU are locked, he cannot change them.

Quote:

---

i had never seen worst mobo ever ...

---

I continue to get the very distinct impression that this product was prematurely released to market... it just sounds too buggy.

Quote:

---

Originally Posted by JumpingJack

I think you are confused ... the multipliers on this CPU are locked, he cannot change them.

---

He probably means lower multipliers. ;)

Very good review, altough I find your conclusions somewhat strange. How can you judge overclocking when the highest your mobo gives is 230HTT? It wont hit 3.6Ghz I'm sure but this is unfair. And why is power consumption 'brutal' if it is only a few watts more under load?

Quote:

---

Originally Posted by v0dka

He probably means lower multipliers. ;)

Very good review, altough I find your conclusions somewhat strange. How can you judge overclocking when the highest your mobo gives is 230HTT? It wont hit 3.6Ghz I'm sure but this is unfair. And why is power consumption 'brutal' if it is only a few watts more under load?

---

Ohhhh, can't change the HTT multipliers... hmmmm sorry.

In terms of the HTT not hitting 3.6 GHz, it is running 3.6 GHz (1800 Mhz x 2, AMD quotes their 'GHz' when they should quote MT/s).

Quote:

---

Originally Posted by v0dka

He probably means lower multipliers. ;)

Very good review, altough I find your conclusions somewhat strange. How can you judge overclocking when the highest your mobo gives is 230HTT? It wont hit 3.6Ghz I'm sure but this is unfair. And why is power consumption 'brutal' if it is only a few watts more under load?

---

OK, if my Phenom can do a better OC, i will very happy ... i am sure this is not ptoblem with CPU, but mobo ... BUT WE NEED take OC of that all platform ... RD790FX + Phenom = No OC today, maybe in few weeks will be better but now No OC!

Most important consumption is in IDLE! Because most of time are computers in IDLE ... think about it!

A litle Babylonian confusion going on here. :D

I meant the cpu multiplier (as you said the cpu multi is locked) and 3.6Ghz cpu clock. The cpu clock probably wont hit 3.6Ghz, but given the results of this review its too early to label Phenom as a non overclackable cpu because the mobo cant go >230htt. Clearly the mobo is limiting the overclock.

In his conclusion OBR says that Phenom has no overclocking. So that struck me as odd. 'Brutal' power consumption also seems overdone, although the difference in load is quite high.

OBR did you give Overdrive a try?

you should be able to squeeze up to 2.7 out of it before the htt walls give strife, but perhaps there are differences between boards.

That's on a 790Fx board mAJORD? Any luck with your Crosshair yet?

Nice test btw OBR!

Sounds like the Gigabyte board isn't so hot with only a 230 HT. Who's to say power consumption numbers aren't being affected by a lousy implementation too. I hear the MSI boards are hitting 500 HT though. Thanks for the effort and review. I hope you have better luck in the future.

Thanks for taking the time and effort to post up OBR...:up:

Quote:

---

Originally Posted by v0dka

A litle Babylonian confusion going on here. :D

I meant the cpu multiplier (as you said the cpu multi is locked) and 3.6Ghz cpu clock. The cpu clock probably wont hit 3.6Ghz, but given the results of this review its too early to label Phenom as a non overclackable cpu because the mobo cant go >230htt. Clearly the mobo is limiting the overclock.

In his conclusion OBR says that Phenom has no overclocking. So that struck me as odd. 'Brutal' power consumption also seems overdone, although the difference in load is quite high.

OBR did you give Overdrive a try?

---

Definitely miscommunication -- to quote a Paul Newman movie "What we have here is a failure to communicate" :) my apologies.

I doubt you will see many hit 3.0 GHz. The > 230HTT may be a limit of the HTT rather than the mobo... as you suggested. So the HTT multi needs to be lowered.

@ORB, I don't read Czech -- could you explain to us if you tried changing the HT multiplier -- can it even be changed. Default is 9, but looking at your CPUID, it appears that the HTT is the culprit....

Yeah lixmark.. on the Ggiabyte.

2 Phenoms in the house ATM and both limited by this "wall"

both the same week, both similar OCing on stock 1.25v. 2.55ishGhz for one, closer to 2.6Ghz for the other prime stable.

Don't know about the Xhair. I don't think it's a case of having luckw ith it, just I need to do a rep Windows install to make it work.

Important to rememeber these are NOT simple CPU upgrades as far as windows is concerned, regardless of BIOS support.

It's the equivlent of a whole chipset (northbridge) change on an Intel system. Be ware you may need to rep windows or at the least, delete all your CPU related :banana::banana::banana::banana: out of device manager before installing

Good points being made.

Quote:

---

Originally Posted by User9498

Sounds like the Gigabyte board isn't so hot with only a 230 HT. Who's to say power consumption numbers aren't being affected by a lousy implementation too. I hear the MSI boards are hitting 500 HT though. Thanks for the effort and review. I hope you have better luck in the future.

---

If you would check Gigabytr thread you would know that this board can hit 430HTT with 5x multi. In other words 2150MHz HTT or 4300MT/s. Problem is with clocking HTT 3.0 Phenom uses but this will be ironed with time...

@mAJORD: Would you do it if you got the time? I'm wondering how the AM2 htlink will do under phenom...

Lixmark, My friend has given me his Asus M2N32 SLI deluxe to try on it. I just don't want to rep or reinstall windows on this machine with the Xhair atm.

thats no prob mAJORD, the m2n32 sli deluxe is almost the same, I'm most interested in AM2 htlink performance :)

Quote:

---

Originally Posted by Lightman

If you would check Gigabytr thread you would know that this board can hit 430HTT with 5x multi. In other words 2150MHz HTT or 4300MT/s. Problem is with clocking HTT 3.0 Phenom uses but this will be ironed with time...

---

Oh, ok. My bad. Guess I have some reading to do.

don't worry user9498 , everyones just a bit testy ATM, so much trolling and nonsense going on around these parts :)

Quote:

---

Originally Posted by OBR

OK, if my Phenom can do a better OC, i will very happy ... i am sure this is not ptoblem with CPU, but mobo ... BUT WE NEED take OC of that all platform ... RD790FX + Phenom = No OC today, maybe in few weeks will be better but now No OC!

Most important consumption is in IDLE! Because most of time are computers in IDLE ... think about it!

---

Yeah here's hoping they get some of this straightened out. Seems they really managed to mess up this release. I blame hector ;) By the time I'm able to upgrade my PC (mid summer 08 earliest) perhaps these issues will be ironed out. Guess we'll see.

BTW, that last part, my PC is rarely at idle so that doesn't matter much to me :p:

ANY review thats show Phenom losing to c2q is in a winrar benchmark, clock for clock, is either fake or done by something who doesn't know what they are doing, so nice try obr, thx for nothing. :shakes:

Quote:

---

Originally Posted by OBR

http://www.pctuning.cz/ilustrace3/ob...500/winrar.png

---

Sorry, but I can´t believe this. I did a full Phenom-Review for a german hardware-magazin and I got way better results:

http://www.planet3dnow.de/photoplog/file.php?n=1003&w=o

Phenom @ 2,4 GHz is way faster than QX6850. ;)

Full review: Phenom 9500 @ Planet 3DNow! (german)

Perhaps something is just inconsistent? OBR didn't get his chip to the same speeds as others either so maybe the gigabyte board doesn't play nice or something right now?

Quote:

---

Originally Posted by MusicIsMyLife

Sorry, but I can´t believe this. I did a full Phenom-Review for a german hardware-magazin and I got way better results:

http://www.planet3dnow.de/photoplog/file.php?n=1003&w=o

Phenom @ 2,4 GHz is way faster than QX6850. ;)

Full review: Phenom 9500 @ Planet 3DNow! (german)

---

Final score depends on lot of stuff indeed (RAM settings, mobo, packed files, cached files, hard disk speed)..., in many tests is Phenom in WinRAR slower then Intel, and you have faster ...that is strange ...
You did worst test ever ... professionals dont measure performance in built-in "WinRAR kBytes benchmark" bacause this test dont say anything about real performance, i tested in seconds ... time is better to compare, then stupid KBytes/s ...

PS. AMD OverDrive works perfectly form me. Everything i changed there, was relly changed in system ... but OC was the same like thru BIOS ...
Now i am waiting for MSI RD-790X and Asus RD-790FX mobos, will be delivered next week. I will test my Phenom on that mobos, and we will see how can overclock this chip on another mobo then strange Gigabyte ...

Quote:

---

Originally Posted by DoubleZero

ANY review thats show Phenom losing to c2q is in a winrar benchmark, clock for clock, is either fake or done by something who doesn't know what they are doing, so nice try obr, thx for nothing. :shakes:

---

We are waiting your reviews done by somebody who do know what he is doing...

It 's fantastic, all the press show phenom isn't as good as we like, now, some have cpus on XS, that you could buy doublezero, that show the same but this is fake....
:down:

Quote:

---

Originally Posted by MusicIsMyLife

Sorry, but I can´t believe this.

---

As I know, Winrar's speed depends on archive contest. For some files you will need more cpu power, for other files you will need more memory bandwidth. So, you can use internal benchmark which mostly tests memory speed or you can use your own set of files.

Quote:

---

Originally Posted by nemrod

We are waiting your reviews done by somebody who do know what he is doing...

It 's fantastic, all the press show phenom isn't as good as we like, now, some have cpus on XS, that you could buy doublezero, that show the same but this is fake....
:down:

---

No, my explanation is above ...

In WinRar are very important : memory speed + timings, pre cached or not pre cached files (try do the same winrar test twice in sequence, you will see massive difference), very important are sort of packed files ... (video, mp3 or pictures and programs) ... every this depends on score in built-in benchmark, OK in this benchmark is Phenom better, but in real usage is slower (in seconds) ...

Quote:

---

Originally Posted by OBR

No, my explanation is above ...

In WinRar are very important : memory speed + timings, pre cached or not pre cached files (try do the same winrar test twice in sequence, you will see massive difference), very important are sort of packed files ... (video, mp3 or pictures and programs) ... every this depends on score in built-in benchmark, OK in this benchmark is Phenom better, but in real usage is slower (in seconds) ...

---

You have a point there... it's just that every credible review that i saw had phenom being faster then C2Q in winrar, it's one of the few benchmarks where K10 wins in desktop.

So my apologies for pointing the finger so quicky :shakes: :D

Hmmm....

Well while these kind of pics and reports are great and all for review articles.

For forums use, I for one would much prefer to see some user results with actual screenshots.... as opposed to pre-compiled graphs and data.....

Lets see some actual screenshots..... thats what I'd prefer to see.... something where you can actaully follow the steps taken by a user during the OC process....

These kind of results are similar to what we see pre-launch, I'm far more interested in seeing what users are able to achieve, with the screenshots and details of whats been tried, to see where problem areas are, or what should be avoided.

Theres no doubt in my mind at the moment that the majority of issues lie, with poor mobo support, and hopefully this will sort itself out in time as mobo manufacturers work out the glitches.....!!

Quote:

---

Originally Posted by kl0012

As I know, Winrar's speed depends on archive contest. For some files you will need more cpu power, for other files you will need more memory bandwidth. So, you can use internal benchmark which mostly tests memory speed or you can use your own set of files.

---

Quote:

---

Originally Posted by OBR

In WinRar are very important : memory speed + timings, pre cached or not pre cached files (try do the same winrar test twice in sequence, you will see massive difference), very important are sort of packed files ... (video, mp3 or pictures and programs) ... every this depends on score in built-in benchmark, OK in this benchmark is Phenom better, but in real usage is slower (in seconds) ...

---

Please, explain this:

For every plattform I install a fresh version of windows xp. I use the same skins, same drivers (only mb-drivers may differ), same settings, same pagefile-size. Before I bench winrar, I reboot the system (not only winrar, for every single benchmark). I open winrar, open the same archive every time (and when I say same archive there should be a high chance to see the same files in archive ;) ) and than I run the benchmark once. When I wan´t to do the benchmark for a 2nd time, I reboot my pc.

Why should there be a difference?

Quote:

---

Originally Posted by OBR

You did worst test ever ... professionals dont measure performance in built-in "WinRAR kBytes benchmark" ...

---

Thank you for these kind words. :rolleyes:

Thank you OBR for your efforts and work :)! As to the whole Winrar benching, I have a solution: do both methods. If ppl care about one over the other, great. Can it seriously hurt to have too much info?

Can't we all just get along? :up:

I'm with MusicIsMyLife on this.No way that C2Q is better in Winrar.Something is wrong in that test done by OBR.

And OBR,you say power consumption is brutal(lol),and in load 9700 draws 11W more (shock => :rolleyes: ).
And in idle,i somehow have a feeling that Phenom with all the advanced power savings integrated in it( better than what C2Q can offer) has actually so much higher idle power cons.This is probably the BIOS/driver issue,since Phenoms can deactivate not only the cores,but IMC too in idle mode.So i vote this out as CnQ2/CoolCore deactivated or not working as they should.

Ok, I didn't read the whole review but Nice mem config in that picture, last I checked that would be single chan wouldn't it 0.0? (on the intel rig) I hope you didn't use this setup for testing :D


http://www.pctuning.cz/ilustrace3/ob...est_system.jpg

double post!

Quote:

---

Originally Posted by kazgirl

Hmmm....

Well while these kind of pics and reports are great and all for review articles.

For forums use, I for one would much prefer to see some user results with actual screenshots.... as opposed to pre-compiled graphs and data.....

Lets see some actual screenshots..... thats what I'd prefer to see.... something where you can actaully follow the steps taken by a user during the OC process....

These kind of results are similar to what we see pre-launch, I'm far more interested in seeing what users are able to achieve, with the screenshots and details of whats been tried, to see where problem areas are, or what should be avoided.

Theres no doubt in my mind at the moment that the majority of issues lie, with poor mobo support, and hopefully this will sort itself out in time as mobo manufacturers work out the glitches.....!!

---

actual screen shots would be believable, i could have took these pictures from any forum and made up the same story :rolleyes:

Quote:

---

Originally Posted by MusicIsMyLife

Please, explain this:

For every plattform I install a fresh version of windows xp. I use the same skins, same drivers (only mb-drivers may differ), same settings, same pagefile-size. Before I bench winrar, I reboot the system (not only winrar, for every single benchmark). I open winrar, open the same archive every time (and when I say same archive there should be a high chance to see the same files in archive ;) ) and than I run the benchmark once. When I wan´t to do the benchmark for a 2nd time, I reboot my pc.

Why should there be a difference?

---

You will see no difference. But if you have two different files with the same size (for example, a text file and a video file), you will see that archiver spent different amount of time for compressing each file, despite the same files size. For a files with a different structures archiver will perform different amount of work. That is, at one of the files will be spent more processing time than at the second and it is becoming more cpu-dependent. Summing up, one can say that for certain types of files, the speed of compression will depend on the speed of memory and for other types of files, the speed will depend on CPU power.
To cite an example of video encoding (which, in a way, is also archiving). Encoding two hours a black background is completed much faster than encoding 2 hours video scenes with the dynamic pursuit.
And it is possible that on cpu-dependent compression Intel quad will be faster then Phenom.

Quote:

---

Originally Posted by informal

No way that C2Q is better in Winrar.

---

Any proof?

Quote:

---

Originally Posted by kazgirl

Hmmm....

Well while these kind of pics and reports are great and all for review articles.

For forums use, I for one would much prefer to see some user results with actual screenshots.... as opposed to pre-compiled graphs and data.....

Lets see some actual screenshots..... thats what I'd prefer to see.... something where you can actaully follow the steps taken by a user during the OC process....

These kind of results are similar to what we see pre-launch, I'm far more interested in seeing what users are able to achieve, with the screenshots and details of whats been tried, to see where problem areas are, or what should be avoided.

Theres no doubt in my mind at the moment that the majority of issues lie, with poor mobo support, and hopefully this will sort itself out in time as mobo manufacturers work out the glitches.....!!

---

:yepp: :clap:

Quote:

---

Originally Posted by OBR

OK, if my Phenom can do a better OC, i will very happy ... i am sure this is not ptoblem with CPU, but mobo ... BUT WE NEED take OC of that all platform ... RD790FX + Phenom = No OC today, maybe in few weeks will be better but now No OC!

Most important consumption is in IDLE! Because most of time are computers in IDLE ... think about it!

---

Who has a computer on IDLE here in Xtremesystems? You should be running F@H or something from the World Community Grid anytime your system is on. :D

BTW...I wonder why the motherboard BIOS's are basically CRAP right now? Hopefully, they'll figure this out before the end of 2007.

Quote:

---

Originally Posted by freeloader

Who has a computer on IDLE here in Xtremesystems? You should be running F@H or something from the World Community Grid anytime your system is on. :D

BTW...I wonder why the motherboard BIOS's are basically CRAP right now? Hopefully, they'll figure this out before the end of 2007.

---

Because AMD undertook the near impossible task of attempting to
deliver all three "Spider" platform components (CPU, MOBO, GPU)
to the Market at the same time.

To accomplish such an (overly?)optimistic goal, it's inevitable that
manufacturers will be pressured to push product out the door early
in order to coordinate the simultaneous launch.

My explanation certainly doesn't justify the rush-to-market, but it
does shed light on the current issues that are putting a "false ceiling"
on current performance.

EBL

Quote:

---

Originally Posted by The-PHENOMATOR

actual screen shots would be believable, i could have took these pictures from any forum and made up the same story :rolleyes:

---

Pics and screenshots are in the link in the first post.

Quote:

---

Originally Posted by kl0012

You will see no difference.

---

Thanks for your comment.



But OBR, I know, I´m stupid. Because I don´t do it the way anybody else does...:rolleyes:

Quote:

---

Originally Posted by MusicIsMyLife

Quote:

---

Originally Posted by kl0012

You will see no difference.

---

Thanks for your comment.



But OBR, I know, I´m stupid. Because I don´t do it the way anybody else does...:rolleyes:

---

Since your previous post, OBR doesn't answer. :D
Are you sure to have read the good number of your winrar test? :ROTF:

Sorry just joking a little bit.