The toy after a week of use

[lopri]

long time, AMD
It's my first AMD CPU in a long time. (not counting the 4050e which came as a combo with the board) I'm excited.

----------------

Note: Below is the last post I made in this thread for anyone interested in the result of my 955BE abuse

I think this will be the last report on my 955BE overclocking. Can't believe it's already been a week since I got this chip. I've learned quite a bit clocking-wise, but the mystery of Phenom is profound. :laugh: (more on that later)

My setup is the #3 rig in my sig. OS is 64-bit Windows 7, and I am utilizing 4x2GB of DDR2-800 RAM (SPD is DDR2-667).

• Max OC using air-cooling (Scythe Infinity): CPU 3.8 GHz (1.39V) / NB 2.6 GHz (1.30V)
  Max CPU OC using two fans (push-pull): CPU 3.9 GHz (1.42V) / NB 2.0 GHz (1.10V)

I have learned that this CPU likes 'just right' amount of voltage, and the NB seems to respond very sensitively to temperatures. (i.e. the lower the better) Towards the end, my choice of the 'sweet spot' for 24/7 OC is CPU 3.6 GHz (1.35V) / NB 2.6 GHz (1.25V). This combination gave me the best blend of performance and thermal characteristics. Another plus is this combo allows me to not raise the core voltage. Stability was tested with many loops of Linpack, PCMark Vantage, 3DMark Vantage, Cinebench, and 4 simultaneous instances of Super pi runs.

Linpack 20 pass Running

Linpack 20 pass Ending

Linpack & 3DMark06

3DMark Vantage

PCMark Vantage

Very satisfying experience, to say the least.

P.S. It has been brought up in this thread, but the way Phenom's IMC/NB works is pretty unique. Not that I figured it out, but I have some data to start with. I ran quick cpu-z latency test with CPU and NB under same frequencies, and different frequencies, then measured the L3 cache latency. The result is intriguing.

CPU -------- NB -------- L3 Latency
1600 ------ 1600 ------ 41 cycles
1800 ------ 1800 ------ 41 cycles
2000 ------ 2000 ------ 41 cycles
..
2600 ------ 2600 ------ 41 cycles

L3 latency stays the same (41 cycles) as long as CPU and NB are a same frequency. Next I fixed the NB frequency @1.6GHz and raised the CPU frequency using multipliers. I don't know how this discrepancy will affect the performance of the CPU, but hopefully I'll find out with others' help.

CPU -------- NB -------- L3 Latency
1600 ------ 1600 ------ 41 cycles
1800 ------ 1600 ------ 42 cycles
2000 ------ 1600 ------ 45 cycles
2200 ------ 1600 ------ 49 cycles
2400 ------ 1600 ------ 51 cycles
2600 ------ 1600 ------ 55 cycles

I don't know what this means, but this is the beginning. Screenshots if you're interested -> Click

 

[Rubycon]

Are you going to overcluck it?

 

[Tempered81]

That processor holds the Quad core OC'ing World record now. Claked at 7.2ghz on two cores and 7.0ghz on all 4 cores.

Good luck - hope to see you take it to at least 4ghz

-oh btw, 47X multiplier if your mobo has the capability.

 

[lopri]

Originally posted by: Rubycon
Are you going to overcluck it?

Was thinking about ovarclucking it.. Over > ovar?

Originally posted by: jaredpace
That processor holds the Quad core OC'ing World record now. Claked at 7.2ghz on two cores and 7.0ghz on all 4 cores.

Good luck - hope to see you take it to at least 4ghz

-oh btw, 47X multiplier if your mobo has the capability.

Ty. There are some disappointments from what I've read (such as lower ovarcluck under 64-bit), but I think it will be fun. It's my first CPU EVAR that has unlocked multipliers! Finally!

P.S. Is the pic @XS you? (under the same nick)

 

[Tempered81]

Congrats on the CPU. I've always wanted an unlocked black edition / extreme edition cpu myself. The pic is good ole B.G.

 

[nyker96]

955BE is excellent, about on par with Q9xxx series. let us know your results.

 

[LOUISSSSS]

write up a guide on how to overclock these things PLEASEEEE

i've NEVER understood how to overclock AMD's and that A64 guide did not help me

 

[BTRY B 529th FA BN]

Looks very interesting

Word to your motha - Vanilla Ice

 

[n7]

w00

 

[SlowSpyder]

:thumbsup:

Good luck, let us know how things go with the new chip.

 

[Rick James]

955 is just a 940 with a multiplier increase correct?

 

[heyheybooboo]

Originally posted by: lopri
long time, AMD
It's my first AMD CPU in a long time. (not counting the 4050e which came as a combo with the board) I'm excited.

at 9 dot 5 dot 5

Good luck - - - - let us know how well that IMC/NB cranks.

Are you going DDR2 or DDR3 ??

 

[SlowSpyder]

Originally posted by: Rick James
955 is just a 940 with a multiplier increase correct?

There are supposed to be some improvements to AMD's 45nm process, most reviewers are able to go further on stock volts it seems when overclocking.

It's more of a higher clocked version of the 945 than the 940. The 940 can only use DDR2. The 945 can use DDR2 or DDR3 depending on what motherboard you stick it in.

But for the most part, from a performance aspect, yea, it's just a higher clocked (by 200MHz) verion of the 945.

 

[jandlecack]

The backwards compatibility to AM2+ seriously holds this chip back. In turn, the DDR3 support doesn't increase performance to warrant AM3+.

Way to screw up both aspects of your processor instead of focussing on one and getting it right, AMD. You can either have a bleeding edge CPU or you can have a compatible mainstream model. Not both in one.

Anyway the other AMD chips currently out aren't even worth looking into, the only choice for a new system if it's AMD is this CPU. Says a lot about their product range, doesn't it?

 

[veri745]

Originally posted by: jandlecack
Anyway the other AMD chips currently out aren't even worth looking into, the only choice for a new system if it's AMD is this CPU. Says a lot about their product range, doesn't it?

You are absolutely correct. I have not heard of anyone that is happy with their purchase of an PhII X3 720.
/s

 

[Rick James]

Originally posted by: jandlecack

Anyway the other AMD chips currently out aren't even worth looking into, the only choice for a new system if it's AMD is this CPU. Says a lot about their product range, doesn't it?

Agreed. X3 720 wasn't that great of a chip.

 

[lopri]

Originally posted by: LOUISSSSS
write up a guide on how to overclock these things PLEASEEEE

i've NEVER understood how to overclock AMD's and that A64 guide did not help me

1. Hit 'Del' at boot and enter BIOS.
3. Change multiplier from 16 to 20.
3. Hit 'F10' then 'Enter'. Wait 30 secs.

1. Shut down and change memory sticks.
2. Rinse and repeat above steps.

Very quick water-testing from last night. Everytime it doesn't like some changes to NB -> clear CMOS. :thumbsdown: Hopefully I will get to run some benches by the weekend.

 

[SlowSpyder]

Originally posted by: jandlecack
The backwards compatibility to AM2+ seriously holds this chip back. In turn, the DDR3 support doesn't increase performance to warrant AM3+.

Way to screw up both aspects of your processor instead of focussing on one and getting it right, AMD. You can either have a bleeding edge CPU or you can have a compatible mainstream model. Not both in one.

Anyway the other AMD chips currently out aren't even worth looking into, the only choice for a new system if it's AMD is this CPU. Says a lot about their product range, doesn't it?

Why do you feel backwards compatibility holds the chip back?

 

[lopri]

I don't think AMD's transition from DDR2->DDR3 is particularly a bad one at this time. Actually I think it helped me choosing a board. (This board is almost like a twin to the P45 board I have). I was seriously considering DDR3, and i7 would have been a logical direction - but I have collected quite a few DDR2 sticks over the years and I couldn't just throw them away. (they're worth nothing now, like DDR2-667 rated stuff) It was by no means the deciding factor but certainly a 'plus' when I made the decision on 955. We'll see if it was the right decision - hopefully soon!

 

[jandlecack]

Originally posted by: SlowSpyderWhy do you feel backwards compatibility holds the chip back?

I'd like to think that by maintaining compatibility of the older socket as well as DDR2 RAM, something had to be sacrificed. This is evident as running the 955 on an AM3-DDR3 platform does not bring the performance gains it should compared to AM2-DDR2, but comes with the price premium.

The DDR3 memory by nature may give it some gains in synthetic benchmarks, but in real-world performance (games anyone) there is almost zero point in running it on AM3 unless you got money to burn and like to boast about running DDR3 or something.

This is probably tied to the lack of triple channel support in the CPU architecture/memory controller. Which brings me back to the compatibility issue...

It's still a very fast chip in its own right though.

 

[SlowSpyder]

Originally posted by: jandlecack

Originally posted by: SlowSpyderWhy do you feel backwards compatibility holds the chip back?

I'd like to think that by maintaining compatibility of the older socket as well as DDR2 RAM, something had to be sacrificed. This is evident as running the 955 on an AM3-DDR3 platform does not bring the performance gains it should compared to AM2-DDR2, but comes with the price premium.

The DDR3 memory by nature may give it some gains in synthetic benchmarks, but in real-world performance (games anyone) there is almost zero point in running it on AM3 unless you got money to burn and like to boast about running DDR3 or something.

This is probably tied to the lack of triple channel support in the CPU architecture/memory controller. Which brings me back to the compatibility issue...

It's still a very fast chip in its own right though.

My guess is they took the DDR2 controller and were able to add on to it, we really have no idea if anything had to be sacrificed. Kind of like when they went with 64 bit support, they built it on to their 32 bit chip.

I think when the transition from DDR1 to DDR2 occured, it was pretty much the same story. DDR2 didn't really offer much over DDR1. My guess is the better timings make up for the bandwidth advantage of DDR3 in a lot of situations. I'm willing to bet as DDR3 improves and as the need for more bandwidth grows DDR3 will start to seperate itself from DDR2.

AMD does lack triple channel support, but does that make a real difference? Again, I wouldn't be suprised if this is a feature that starts to show it's true benefit down the road, but not sure how much it matters right now, or if it really makes a point that AMD had to sacrifice anything for backwards compatibility.

Just my $.02.

 

[Flipped Gazelle]

Originally posted by: jandlecack

Originally posted by: SlowSpyderWhy do you feel backwards compatibility holds the chip back?

I'd like to think that by maintaining compatibility of the older socket as well as DDR2 RAM, something had to be sacrificed. This is evident as running the 955 on an AM3-DDR3 platform does not bring the performance gains it should compared to AM2-DDR2, but comes with the price premium.

The DDR3 memory by nature may give it some gains in synthetic benchmarks, but in real-world performance (games anyone) there is almost zero point in running it on AM3 unless you got money to burn and like to boast about running DDR3 or something.

This is probably tied to the lack of triple channel support in the CPU architecture/memory controller. Which brings me back to the compatibility issue...

It's still a very fast chip in its own right though.

I'm still not sure what you are getting at. No transition in memory standard since SDR-DDR has done much, if anything, for performance. In fact when DDR2 was introduced, it was found to be slower than DDR-400.

From your post, it sounds like you are making suppositions based on zero knowledge.

 

[jandlecack]

Originally posted by: SlowSpyder

Originally posted by: jandlecack

Originally posted by: SlowSpyderWhy do you feel backwards compatibility holds the chip back?

I'd like to think that by maintaining compatibility of the older socket as well as DDR2 RAM, something had to be sacrificed. This is evident as running the 955 on an AM3-DDR3 platform does not bring the performance gains it should compared to AM2-DDR2, but comes with the price premium.

The DDR3 memory by nature may give it some gains in synthetic benchmarks, but in real-world performance (games anyone) there is almost zero point in running it on AM3 unless you got money to burn and like to boast about running DDR3 or something.

This is probably tied to the lack of triple channel support in the CPU architecture/memory controller. Which brings me back to the compatibility issue...

It's still a very fast chip in its own right though.

My guess is they took the DDR2 controller and were able to add on to it, we really have no idea if anything had to be sacrificed. Kind of like when they went with 64 bit support, they built it on to their 32 bit chip.

I think when the transition from DDR1 to DDR2 occured, it was pretty much the same story. DDR2 didn't really offer much over DDR1. My guess is the better timings make up for the bandwidth advantage of DDR3 in a lot of situations. I'm willing to bet as DDR3 improves and as the need for more bandwidth grows DDR3 will start to seperate itself from DDR2.

AMD does lack triple channel support, but does that make a real difference? Again, I wouldn't be suprised if this is a feature that starts to show it's true benefit down the road, but not sure how much it matters right now, or if it really makes a point that AMD had to sacrifice anything for backwards compatibility.

Just my $.02.

You make some valid points, and you are correct about your DDR1->DDR2 analysis of course. The lower latencies of the established DDR1 allowed it to render DDR2 platforms insignificant for some time, due to early DDR2's higher latencies and small frequency increments.

However, with DDR3 we are already seeing mighty fast frequency specifications at 1.65V, which as part of the equation make the comparably higher latencies not an issue. As with the review you posted, and correctly assumed, as DDR3 speed increases, it makes more and more difference to run it in triple channel over dual channel. It is however also true that triple channel PC10600 vs. PC12800 or - if memory serves correctly - even PC16000 for example does not give a noticeable improvement in real-world applications. Therefore, the biggest factor, albeit not the only one, when picking a new platform and comparing RAM will become the dual channel vs triple channel, since triple vs triple does not set itself apart very well at the moment anyway.

This is where I see the current PHII architecture falling behind a little. I don't think many RAM sales will come from higher clocked sticks, but rather the low-end DDR3 models, because as I mentioned previously the clock-for-clock in same channel comparison doesn't throw up significant results.

Originally posted by: Flipped Gazelle

Originally posted by: jandlecack

Originally posted by: SlowSpyderWhy do you feel backwards compatibility holds the chip back?

I'd like to think that by maintaining compatibility of the older socket as well as DDR2 RAM, something had to be sacrificed. This is evident as running the 955 on an AM3-DDR3 platform does not bring the performance gains it should compared to AM2-DDR2, but comes with the price premium.

The DDR3 memory by nature may give it some gains in synthetic benchmarks, but in real-world performance (games anyone) there is almost zero point in running it on AM3 unless you got money to burn and like to boast about running DDR3 or something.

This is probably tied to the lack of triple channel support in the CPU architecture/memory controller. Which brings me back to the compatibility issue...

It's still a very fast chip in its own right though.

I'm still not sure what you are getting at. No transition in memory standard since SDR-DDR has done much, if anything, for performance. In fact when DDR2 was introduced, it was found to be slower than DDR-400.

From your post, it sounds like you are making suppositions based on zero knowledge.

Yes, DDR2 was found slower or equal to DDR1 at first, but this isn't the case with DDR3 vs DDR2 anymore. Unless price is an issue, any new system should be built for DDR3 in mind, which is why the lack of triple channel is inexcusable in my opinion. Were you making a point somewhere, or would you prefer to fail at trying to judge my computing related background?

 

[Flipped Gazelle]

Originally posted by: jandlecack

Originally posted by: SlowSpyder

Originally posted by: jandlecack

Originally posted by: SlowSpyderWhy do you feel backwards compatibility holds the chip back?

I'd like to think that by maintaining compatibility of the older socket as well as DDR2 RAM, something had to be sacrificed. This is evident as running the 955 on an AM3-DDR3 platform does not bring the performance gains it should compared to AM2-DDR2, but comes with the price premium.

The DDR3 memory by nature may give it some gains in synthetic benchmarks, but in real-world performance (games anyone) there is almost zero point in running it on AM3 unless you got money to burn and like to boast about running DDR3 or something.

This is probably tied to the lack of triple channel support in the CPU architecture/memory controller. Which brings me back to the compatibility issue...

It's still a very fast chip in its own right though.

My guess is they took the DDR2 controller and were able to add on to it, we really have no idea if anything had to be sacrificed. Kind of like when they went with 64 bit support, they built it on to their 32 bit chip.

I think when the transition from DDR1 to DDR2 occured, it was pretty much the same story. DDR2 didn't really offer much over DDR1. My guess is the better timings make up for the bandwidth advantage of DDR3 in a lot of situations. I'm willing to bet as DDR3 improves and as the need for more bandwidth grows DDR3 will start to seperate itself from DDR2.

AMD does lack triple channel support, but does that make a real difference? Again, I wouldn't be suprised if this is a feature that starts to show it's true benefit down the road, but not sure how much it matters right now, or if it really makes a point that AMD had to sacrifice anything for backwards compatibility.

Just my $.02.

You make some valid points, and you are correct about your DDR1->DDR2 analysis of course. The lower latencies of the established DDR1 allowed it to render DDR2 platforms insignificant for some time, due to early DDR2's higher latencies and small frequency increments.

However, with DDR3 we are already seeing mighty fast frequency specifications at 1.65V, which as part of the equation make the comparably higher latencies not an issue. As with the review you posted. and correctly assumed, as DDR3 speed increases, it makes more and more difference to run it in triple channel over dual channel. It is however also true that triple channel PC10600 vs. PC12800 or - if memory serves correctly - even PC16000 for example does not give a noticeable improvement in real-world applications. Therefore, the biggest factor, albeit not the only one, when picking a new platform and comparing RAM will become the dual channel vs triple channel, since triple vs triple does not set itself apart very well at the moment anyway.

This is where I see the current PHII architecture falling behind a little. I don't think many RAM sales will come from higher clocked sticks, but rather the low-end DDR3 models, because as I mentioned previously the clock-for-clock in same channel comparison doesn't throw up significant results.

But just because tripple channel works well with i7's architecture, does not mean that it will work well the PhII's. Remember, PhII is very much a legacy architecture, it's roots deep in the K8 reservoir. It's merely evolutionary, not revolutionary.

 

[jandlecack]

Originally posted by: Flipped Gazelle
But just because tripple channel works well with i7's architecture, does not mean that it will work well the PhII's. Remember, PhII is very much a legacy architecture, it's roots deep in the K8 reservoir. It's merely evolutionary, not revolutionary.

My point exactly.

The architecture is not as impressive as it should be, considering it's AMDs flagship model, and mostly the only reason to build an AM3 based system.