Could B3 Phenom be the rebirth of AMD....+1000MHZ OC done!

[lopri]

Cinebench 9.5 should be quick and easy. Thanks.

A Q: Are you manually setting the NB clock or does it proportionally clock to main clock? In either case, is there a limit how high NB clock will go? (e.g. At CPU xxxxMHz the max achievable NB is xxxxMHz, etc.)

I noticed the read bandwidth of L3 was similar to memory bandwidth in your Everest shot - do they move up/down together?

[gOJDO]

super pi...man i said like 10 times i won't be running it

I might have missed that part, but why not run Super Pi?
As we know, it is a pure single-threaded application and it's performance depends of the cache size and speed. Also we know that the L3 on K10 can be dedicated to only 1 core if needed and that's exactly what will happen while running Super Pi. Let's see by how much the NB & L3 speed can boost the performance of K10 in single-threaded apps.

3d01 is ok but single thread

Exactly, that's why I put it in the list.

winrar (is it multi thread now?)

Yes, since 3.70.

Thanks

[Piotrsama]

Good work Tony, keep it up.

[afireinside]

Yes they look for fuel to bring down the thread with why run this at this speed when c2d runs this etc...im not interested

I must be the minority then, I just want to see how a properly tweaked phenom compared to K8. The 3d clocks are starting to get good. I bet that in some weeks the retail chips will be flying.

[Tony]

I must be the minority then, I just want to see how a properly tweaked phenom compared to K8. The 3d clocks are starting to get good. I bet that in some weeks the retail chips will be flying.

I bought this CPU from www.aria.co.uk is this not a retail CPU ?

I can run a 6400+ also although many have posted results already with these cpu's

[dinos22]

Just been Googling how popular this thread is, seems a few are thinking I have been running massive voltages into the Phenom....sorry guys 1.5V max here for the 3.5GHZ and everything else between 1.3 and 1.45V Default is 1.3V.

seems cpu-z etc has been reading a little high.

the level of trust is some apps just amazes me...

you should make that clear in the first post or ppl only have that info to go on

CPUz generally reads vcore correctly
at least on most intel platforms other than some noteable ones like Foxconn boards and possibly a few others

[akaBruno]

Tony:

Just keep up the good work and SCREW THE TROLLS!!!

They'd like nothing more than seein this thread end.

They're trying to bait you.

Bruno

[Brother Esau]

Tony:

Just keep up the good work and SCREW THE TROLLS!!!

They'd like nothing more than seein this thread end.

They're trying to bait you.

Bruno

Agreed and unfortunately true because we see it in every single AMD thread on the forum and those that be turn a blind eye to it

[Tony]

you should make that clear in the first post or ppl only have that info to go on

CPUz generally reads vcore correctly
at least on most intel platforms other than some noteable ones like Foxconn boards and possibly a few others

you know I think I know why it was reading high, I use vcore+ special add on the DFI, not the VIDS, maybe this was why the readings were screwed.

overall though I NEVER trust apps that read voltages off a board, if knowing the voltage is important you should always measure on the board to be sure.

I have seen differences so huge between a DMM and bios etc that you would think the board was faulty...even 12V lines showing 11.2V in bios and windows that were at 12.3V real.

[i found nemo]

they are happy cuz they FINALLY get to talk crap to amd users after us talking crap to them for ... how many years? but enough talk about trollers. tony i'd like to see you run cinebench and maybe a crysis bench if attainable. cpu mark 2

http://www.softpedia.com/get/System/.../CPUMark.shtml

and cpu mark 1.0

[iadstudio]

Wow! Great work Tony. Please link to your article when it's up.

[dinos22]

you know I think I know why it was reading high, I use vcore+ special add on the DFI, not the VIDS, maybe this was why the readings were screwed.

overall though I NEVER trust apps that read voltages off a board, if knowing the voltage is important you should always measure on the board to be sure.

I have seen differences so huge between a DMM and bios etc that you would think the board was faulty...even 12V lines showing 11.2V in bios and windows that were at 12.3V real.

agree with you mate

i always measure against software readings and if i see they are same or very very close i rely on software otherwise will stick with DMMs

[afireinside]

I bought this CPU from www.aria.co.uk is this not a retail CPU ?

I can run a 6400+ also although many have posted results already with these cpu's

Worded poorly, I meant later on into production as they mature.

[SocketMan]

Here is the "conclusion" of an excellent analysis between k8/k10.

http://www.digit-life.com/articles3/...ma-phenom.html

See above for full version, it's all in bytes/instructions per cycle and
nanoseconds so (if you need to) it's "relatively" easy to calculate performance for given (say 3.5 ) Mhz level.

Quote:

The architecture of new AMD Phenom processors (AMD K10) reviewed in this article is quite successful. Its main advantage over the previous architecture (AMD Athlon 64, K8) is its balanced design on all levels. For example, the L1 D-Cache bus is expanded to 256 bit (2x128) agrees well with 128-bit FP arithmetic units, providing equal bandwidth for reading data using 128-bit SSE/SSE2 multimedia instructions (32 bytes/cycle). And the 128-bit FP units per se can yield a noticeable performance gain in most synthetic and real applications working with this data type (e.g. scientific programs, 3D Modeling, computer games, etc). In our simple case (reading data with 128-bit SSE/SSE2 instructions), their execution speed has grown fourfold.

Expanded to 128 bit in one direction, the L1-L2 D-Cache bus (which organizational type is still in question) eliminates a bottleneck in the data exchange process between a processor core and system memory (it's hard to imagine that the L1-L2 bus is as wide as the L2-memory controller bus in AMD K8 processors!) A great advantage of the new architecture is its efficient hardware prefetch on all levels (L2, L3 Caches, and system memory). As we have shown in our separate review, it can show its true worth in case of multi-core access to system memory, reaching up to 85% of performance potential of high speed dual-channel DDR2 memory.

What concerns L3 Cache in its integrated memory controller shared for all processor cores, this cache level is not an apparent advantage of new processors. According to our single-core access tests, bandwidth of this cache level is not much higher than bandwidth of system memory in the same conditions (8.9 GB/s versus 6.4 GB/s), which has to do with high bandwidth of dual-channel DDR2 memory. L3 Cache has better latencies, but only for random memory access (in other cases latencies of all levels are effectively "camouflaged" by hardware prefetch). So its advantages can be more noticeable in applications that randomly access data stored in different areas of memory (for example, databases).

End quote

[wiak]

LOL...i wish

its fun to tweak AMD's CPU's, I get bored of tRD etc on intel.

I wonder how many I sold today ?

dont know i ordered 2 days ago, should get mine today
tony Gigabyte MA-790FX-DS5 should be as good as DQ6, same board, just missing some network, pcie slots

i would LOVE some x264 benchmark!!
http://www.techarp.com/showarticle.aspx?artno=442

great work amd
Tony what is the max oc using "auto" settings in the gigabyte bios/board?
hehe

[wiak]

dang it, another life sucking thread i'm going to have to follow....

agreed
+1

[wolf2009]

u have made me think highly of amd again . u have made me come to this amd thread on this forum . thank u .

[JumpingJack]

super pi...man i said like 10 times i won't be running it

3d01 is ok but single thread, winrar (is it multi thread now?) HD encoding...I would need a HD source to re encode wouldn't I?

CS2/3...is that free..i think not

DivX...i could re encode a tv cap here from mpg.

http://www.techarp.com/showarticle.aspx?artno=442

With it's own source file. Avisynth is also free. Phenom does well on this. Multithreaded.

[JumpingJack]

i dunno i will give you my user name and password and you can get it from steam... i mean, i trust you tony

This would not be a good idea, it violates the TOS:

You may not reveal, share or otherwise allow others to use your password or Account. You agree that you are personally responsible for the use of your password and Account and for all of the communication and activity on Steam that results from use of your login name and password. You may not sell or charge others for the right to use your Account, or otherwise transfer your Account.

I suspect if they register logons from several different IPs, you may get an email question as to where those IPs reside ... if you answer incorrectly... no more Steam for you ....

Tony can get most any game via a demo download for benching if he so desires.

[Morais]

wPrime Please

[Daveburt714]

Give me few days... ok, few hours ;]

Any results from those bad boys yet ryba?

[Lightman]

I think Quake 3 Arena Timedemo should show L3 cache influence on overlay performance. As we know this engine is very sensitive to memory subsystem.

You can get free demo here: http://www.idsoftware.com/games/quak...e_section=demo

Thanks!

[wiak]

I always look forward to your threads tony, Im learning alot about phenom by reading them.


Anyone else have a 9850 or 9550 in the mail?

i have a 9850 blacky editon and the Gigabyte MA-790FX-DS5 aka little brother of DQ6

He posted Cinebench and 3DMark06 CPU score. I can easily compare the results with Intel stuff. (Of course keeping in mind that AMD scales better in Cinebbech, and that Tony hasn't had a lot of time to play with the CPU)

Can't you?

Or does 'comparison' mean Super Pi by definition? I can guess that part, too. Even @3.5GHz Phenom won't likely break 20 secs/1M. There, you have it. If you want the best Pi times, you should look into Intel CPUs and motherboards. Phenom sucks in Pi calculation and I don't think that's too hard to figure out.

Now, let's get this Pi-this Pi-that out of this thread please. If that's only thing that matters, wait a few days - like Tony says, perhaps someone else will do it.

Personally, I no longer even install the stupid benches. Most of the time I don't find the scores indicative of gaming performance, and I really would rather be playing games than watching my computer do endless runs of a benchmark program. Show me a CPU-Z of it at 3.5 and running a game flawlessly, and I'm perfectly content. Because, in the end, that's why I buy this stuff and abuse it. That, and to feed my case modding obsession.

I understand that not everyone shares my view. I wouldn't want them to. Different viewpoints make this place fun, but just remember that my viewpoint has a place just as well as the superpi supporters.

And as for not benching, you know, if he has limited time I'd rather see him pull up 3.5 to stable than waste time on spi and settle for a slower clock.

what i have heard is that SuperPi is dependant on cache, and that intel cpus have like 50% more cache than amds, that kinda kills the comparison
enough with this SuperPi, i want to see REAL results like x264 video encoding, thats what you mainly use a quad core for
and i do agree with tony, intel comperisons should be NOT in this thread

best wishes to Tony for his phenomenal water adventures

[Morais]

lol intel have more cache because the lack of integrated memory controler on cpu.If they got it, theres no need for exorbitant caches sizes & performance would be better

[Rapt0r]

so is it possible for this baby to hit 3Ghz on stock volts?