Sign in with
Sign up | Sign in
Your question

Phenoms B3 coming in days?? or what????

Tags:
Last response: in CPUs
Share
March 10, 2008 9:38:53 AM

So I havent built yet, been waiting on the phenoms B3 but Im confused on exactly when they are dropping. Im reading everywhere it says march but some saying its now june??? Some other people are saying it may be dropping within days??? Im on an old pentium 4 2.8 ghz socket 478 that I have as an emergency back up from back in the days that I found. So does anyone have any real idea or have an inkling if its gonna drop in days or months??? Also does anyone have an estimated price of the chip? Thanks.

More about : phenoms coming days

March 10, 2008 10:22:24 AM

mrgoodbar said:
So I havent built yet, been waiting on the phenoms B3 but Im confused on exactly when they are dropping. Im reading everywhere it says march but some saying its now june??? Some other people are saying it may be dropping within days??? Im on an old pentium 4 2.8 ghz socket 478 that I have as an emergency back up from back in the days that I found. So does anyone have any real idea or have an inkling if its gonna drop in days or months??? Also does anyone have an estimated price of the chip? Thanks.

Last report had them in May to June. Thats about the same time as HD 4000 so few wires could be crossed. The price of the phenom's at speeds 2.3 and 2.4 should be the same price as current. The 9700 and 9900 price should be cheaper than the Q6700.

I wonder when AMD will put 2 45nm phenom's together for an octal?
March 10, 2008 11:18:20 AM

There was something on google finance about b3 barcelona opterons being ready to launch this week, so unless there's some kind of NDA on them they might at least sneak their way to reviewers within the next couple of weeks.

Either way B3 Phenoms on the shelf and ready to buy will be end of next month I would think. I assume you've been waiting a while to upgrade but in for a penny in for a pound you might as well hold on a bit longer to at least say what the reviews of them say. If the B3s oc better (i.e. say to 2.8ghz) then they might make a reasonable gaming setup, with good non-gaming performance (windows, file decoding & compression).

The new AMD motherboards look pretty sharp I might add, but then again there are loads of quality Intel boards around too...
Related resources
March 10, 2008 2:43:54 PM

spoonboy said:
....

Either way B3 Phenoms on the shelf and ready to buy will be end of next month I would think.

...


Hi

I think you will find the OEMs who weren't prepared to use the buggy B2 stepping will snap up all the B3's first. Please refer to THG interview with Jochen Polster, CEO of AMD Germany!!

Bob
March 10, 2008 3:25:59 PM

Its coming in Q2, with higher speed grade (9900 @2.6Ghz) coming in Q3. As elbert said, both 9900 and 9700 should be cheaper than Q6700, which is scheduled to drop to 260USD (current Q6600 price point). Q6600 will continue drop to 220USD~240USD.
a c 127 à CPUs
March 10, 2008 4:36:29 PM

yomamafor1 said:
Its coming in Q2, with higher speed grade (9900 @2.6Ghz) coming in Q3. As elbert said, both 9900 and 9700 should be cheaper than Q6700, which is scheduled to drop to 260USD (current Q6600 price point). Q6600 will continue drop to 220USD~240USD.


Yea thats the plan. I would say I am sad due to the price cut but still $275 is not bad for one of the best quad cores available.

Lets see what happens with Phenom. I don't expect anything great. Nor do I expect much more in the arena of OC'ing. But we shall see.
March 10, 2008 6:36:46 PM

Alright thanx for the Info.
a c 127 à CPUs
March 10, 2008 11:05:05 PM

I am sorry to report this but Phenom has been delayed again. They are having trouble ramping B3 up and would like you to know they appreciate your patients. It should be ready by Q308 at the latest.







:p  JUST KIDDING......
March 10, 2008 11:44:08 PM

I can NOT wait until the B3 comes out, I hate this E6750. AMD forever!
March 11, 2008 1:01:17 AM

The_Blood_Raven said:
I can NOT wait until the B3 comes out, I hate this E6750. AMD forever!


Eh?
March 11, 2008 11:45:12 AM

The_Blood_Raven said:
I can NOT wait until the B3 comes out, I hate this E6750.
:heink:  Why? Because it runs software? Because it has no AMD logo? Because it doesn't come in designer colors?
The_Blood_Raven said:
AMD forever!
AMD Forever! :pfff: 
Intel Inside! :pfff: 
Best product gets the dollars. Anything else is silly, IMO.
March 11, 2008 12:14:55 PM

I am just happy with my Q6600. So what that it is a couple of dollars more than the Phenom. It’s not buggy, faster in applications the general people use and well, it does not have issues. The best part is, I got mine a year ago and its still better then the Phenom.

What’s not to like?
March 11, 2008 12:50:20 PM

spaztic7 said:

What’s not to like?


Mine gets quite "boggy" when running many apps.


I think its prob the FSB, but can't say for sure...
March 11, 2008 1:06:30 PM

Amiga500 said:
Mine gets quite "boggy" when running many apps.


I think its prob the FSB, but can't say for sure...


It's most likely how much ram you have and how it's setup.
March 11, 2008 1:11:30 PM

deranged said:
It's most likely how much ram you have and how it's setup.


Its got 4GB of ram, and the apps use alot of it, but never dipping into virtual mem (not that I could since only 3.25 registers).
a c 127 à CPUs
March 11, 2008 1:11:47 PM

spaztic7 said:
I am just happy with my Q6600. So what that it is a couple of dollars more than the Phenom. It’s not buggy, faster in applications the general people use and well, it does not have issues. The best part is, I got mine a year ago and its still better then the Phenom.

What’s not to like?


Agreed. Plus a nice 3GHz OC with no voltage change and it roxors the soxors.

Amiga500 said:
Mine gets quite "boggy" when running many apps.


I think its prob the FSB, but can't say for sure...


Eh? Man what are you doing? Running 10 Folding@Homes? I once ran 3 instances of one game, and 2 other different game(all MMORPGs) had WMP plaing music in the background, had DVDShrink running, had a defrag going, had 15+ IE7's and was downloading 2 or 3 things and it never once slowed down. And this is on Vista using only 2GB of RAM. You must not have a Q6600.
March 11, 2008 1:16:15 PM

jimmysmitty said:
Eh? Man what are you doing? Running 10 Folding@Homes? I once ran 3 instances of one game, and 2 other different game(all MMORPGs) had WMP plaing music in the background, had DVDShrink running, had a defrag going, had 15+ IE7's and was downloading 2 or 3 things and it never once slowed down. You must not have a Q6600.


I have and it bogs down when faced with 2 cfx pre processors and 2 cfx post processors along with an icem mesh generator.


Like it or lump it, the fsb is still a bottleneck for some.



Far far better than the old P4 piece of **** I had before, and it is better than the 4600 X2 I have at home.
a c 99 à CPUs
March 11, 2008 1:42:27 PM

Amiga500 said:
Mine gets quite "boggy" when running many apps.


I think its prob the FSB, but can't say for sure...


Probably not, at least from what most of the tests I've seen run here and other places have said. Look at specFP_rate for 1 thread -> 2 thread scaling as that hits the memory I/O about as hard as anything else I've seen. The FSB only starts to be an issue with synthetic benches such as specFP_rate when running an MCM quad-core CPU on an FSB (such as the Q6600) and only starts to be much of an issue with real programs when running two quad-core MCM CPUs on a single system (such as the E53xx series.)

I'd put your bogging down under running many apps as being the fault of the OS (particularly if you're running Windows XP), only having two cores and >2 intensive threads, amount of RAM, HDD I/O and network I/O ahead of FSB congestion on your E6750.
March 11, 2008 3:04:46 PM

spongebob said:
AMD Forever! :pfff: 
Intel Inside! :pfff: 
Best product gets the dollars. Anything else is silly, IMO.


COMPLETELY AGREE.
March 11, 2008 3:10:35 PM

Amiga500 said:
Like it or lump it, the fsb is still a bottleneck for some.


From what I have read here and on other websites regarding intels fsb and bottlenecks, there is no significant increase in speed from higher fsb speeds. A net increase of .05 with 1333 verses 1066 in real world applications. I wish I had the link, as I read it about a half a year ago.

You will see a difference in sinthetic benchmarks- (in the form of memory bandwidth)

As for real world apps, there is a difference in the server market... That's part of the reason AMD has lasted all this time with a slower processor.

AMD's integrated memory controler can acces memory faster and more efficiently, and in the server market where access of a ton of ram is important, the fsb bottleneck manifest itself.

For the adverage joe like you and me, it's pretty tough if not impossible to fully saturate the fsb with a core2.
March 11, 2008 4:03:20 PM

Amiga500 said:
Like it or lump it, the fsb is still a bottleneck for some.



So Intel uses an FSB, and that's a bottleneck.

AMD uses an IMC and according to your thinking isn't bottlenecked.

Then why does Core 2 outperform K10?

Is there a bottleneck at the end of one of the HT links called a "Phenom Processor"?
March 11, 2008 4:15:10 PM

spaztic7 said:
I am just happy with my Q6600. So what that it is a couple of dollars more than the Phenom. It’s not buggy, faster in applications the general people use and well, it does not have issues. The best part is, I got mine a year ago and its still better then the Phenom.
What’s not to like?

jimmysmitty said:
Agreed. Plus a nice 3GHz OC with no voltage change and it roxors the soxors.

^Agreed. My quad is so much faster than my old P4 and overclocked so easily. If she was a girl I would....well I like mine alot. :) 
a c 127 à CPUs
March 11, 2008 4:16:47 PM

ImajorI said:
^Agreed. My quad is so much faster than my old P4 and overclocked so easily. If she was a girl I would....well I like mine alot. :) 


LMFAO!!!!!!!!!! :na:  :lol:  :na:  :lol: 
March 11, 2008 4:37:37 PM

TechnologyCoordinator said:
So Intel uses an FSB, and that's a bottleneck.

AMD uses an IMC and according to your thinking isn't bottlenecked.

Then why does Core 2 outperform K10?

Is there a bottleneck at the end of one of the HT links called a "Phenom Processor"?


LMAO!!!!!!

Everything is good for its day.... Phenoms day was just a long time ago. If not for the delays!
March 11, 2008 4:51:13 PM

MU_Engineer said:
I'd put your bogging down under running many apps as being the fault of the OS


Possible. I'll not discount it.



Although when running the solver on 4 nodes (4 cores effectively), there is a 15% speedup by ensuring the cache isn't wiped by the OS setting proc infinity.

That is a result of keeping stuff away from the FSB as much as possible.




As regards the Barcelona. Yeah, on the desktop its ****. But it would be quicker for me doing what I'm doing - pity AMD didn't get the finger out last summer/autumn when I had to get a new machine or lose my sanity (and its questionable whether I've kept it or not :D  )
March 11, 2008 4:53:42 PM

deranged said:
From what I have read here and on other websites regarding intels fsb and bottlenecks, there is no significant increase in speed from higher fsb speeds. A net increase of .05 with 1333 verses 1066 in real world applications. I wish I had the link, as I read it about a half a year ago.

You will see a difference in sinthetic benchmarks- (in the form of memory bandwidth)

As for real world apps, there is a difference in the server market... That's part of the reason AMD has lasted all this time with a slower processor.

AMD's integrated memory controler can acces memory faster and more efficiently, and in the server market where access of a ton of ram is important, the fsb bottleneck manifest itself.

For the adverage joe like you and me, it's pretty tough if not impossible to fully saturate the fsb with a core2.


I think the FSB being a bottle neck point is more something that would impact a quad-core intel more than a dual-core one. Its a bit long in the tooth, but in the quad cores the fsb is also used for the two dies to talk to each other. The idea is that thats alot of traffic along a route that isn't known for its performance. Therefore FSB increases might enable higher CPU performance than a higher clock with a higher multiplier. Say a q6600 @3.0ghz - 333mhz x 9 would outperform outperform a qx700 (unlocked multiplier) @3.0ghz - 272mhz x 11, due to a bottlenecking of the fsb under heavy load.
March 11, 2008 4:59:09 PM

What is the "Ratio"

People who actually use Quad?
People who just have Quad?

just a little fuggen curious here
March 11, 2008 6:00:22 PM

spoonboy said:
I think the FSB being a bottle neck point is more something that would impact a quad-core intel more than a dual-core one. Its a bit long in the tooth, but in the quad cores the fsb is also used for the two dies to talk to each other. The idea is that thats alot of traffic along a route that isn't known for its performance. Therefore FSB increases might enable higher CPU performance than a higher clock with a higher multiplier. Say a q6600 @3.0ghz - 333mhz x 9 would outperform outperform a qx700 (unlocked multiplier) @3.0ghz - 272mhz x 11, due to a bottlenecking of the fsb under heavy load.



If that's true then like a prev. poster said, wouldn't a similar clocked AMD chip be faster than core2 because of what you think is a fsb bottleneck? The way you put it is the intel fsb bottle necks how the two cpu's communicate in quadcore intels. So by that logic you would think Amd at the same clock rate would be faster, not the case.

If Intel is faster regardless of the more effient architecture, then you have to think the fsb is doing a good job, or the better architecture would be bottlenecked, effectively negating any potential performance gains.

I also want to add that the older extreme edition chips run at 3.0ghz on the same fsb as the q6600. Do all you fsb bottleneckers think intel would sell a 1,000 chip that can't handle it's own data path?

March 11, 2008 6:44:08 PM

Agreed - if the curent quad cores are bottlenecked for efficiency, and its still the fastest out there, i can't wait to see what happens when they move the memory controller to on chip (socket 603/4) whatever its name is.
March 11, 2008 6:49:17 PM

"If that's true then like a prev. poster said, wouldn't a similar clocked AMD chip be faster than core2 because of what you think is a fsb bottleneck? The way you put it is the intel fsb bottle necks how the two cpu's communicate in quadcore intels. So by that logic you would think Amd at the same clock rate would be faster, not the case."

Core 2 is simply better than K8. The FSB might present an inefficiency, but Intel's architecture is more than good enough to shine through an aging interconnect.

The FSB is old, but it's enough at this level, however moving forward Intel will be replacing it with something akin to the Hyper Transport system with it's next generation of processors, so they are acknowledging that it's approaching it's useful limit.

"The way you put it is the intel fsb bottle necks how the two cpu's communicate in quadcore intels"

Easy now, it was just the theory, and I just wrote an example.

"If Intel is faster regardless of the more effient architecture, then you have to think the fsb is doing a good job"

yes you would, and yes I do. My next upgrade I have planned is a q6600 to replace my e6300.

"I also want to add that the older extreme edition chips run at 3.0ghz on the same fsb as the q6600. Do all you fsb bottleneckers think intel would sell a 1,000 chip that can't handle it's own data path?"

No I dont. I used the extreme edition chip as an example because the multiplier can be changed on them to enable the example scenario I layed out. That doesn't necessarily mean though that a higher FSB with a lower multiplier wouldn't be slightly, or theoretically faster.








March 11, 2008 7:21:10 PM

Spoonboy

Come together with your hands
Save me, I'm together with your plan
Save me

March 11, 2008 7:29:15 PM

1801227,30,299418 said:
The FSB is old, but it's enough at this level, however moving forward Intel will be replacing it with something akin to the Hyper Transport system with it's next generation of processors, so they are acknowledging that it's approaching it's useful limit.quotemsg]

We are debating the current fsb with core2 not what is going to happen in the future. (I thought)

Intels road map states that they will be going with in integrated memory controller.

To all that might have thought I was attacking them I'm sorry.. However I'm getting a little sick of people saying Amd or Intel is running slow because of "x" reason without proof or actually testing their claims.

An example is I tested my 1066 ddr2 Crucials at different speeds and latency, as well as at different fsb speeds to see how much of a performance impact I could get. The reason I did this was because so many people were saying that 1:1 ratio is the fastest ratio and anything eles would negate the speed of the ram I just spent big money on. In my case ddr2 1066.

The result was higher memory bandwidth and lower latency in spite of what the fsb was doing.

If I listened to most of the people I would have fast ram running slower than it can right now.
March 11, 2008 8:59:09 PM

^^ Thanks Im gonna wait the 2 weeks and see what happens, this should be an intresting april.
March 12, 2008 2:03:58 PM

If we hold our breath for it, we will all die. How many times has Barcelona been pushed back? How is this any different?
March 13, 2008 2:21:14 PM

deranged said:
1801227,30,299418 said:
The FSB is old, but it's enough at this level, however moving forward Intel will be replacing it with something akin to the Hyper Transport system with it's next generation of processors, so they are acknowledging that it's approaching it's useful limit.quotemsg]

We are debating the current fsb with core2 not what is going to happen in the future. (I thought)

Intels road map states that they will be going with in integrated memory controller.

To all that might have thought I was attacking them I'm sorry.. However I'm getting a little sick of people saying Amd or Intel is running slow because of "x" reason without proof or actually testing their claims.

said:


U spoon head.

"We are debating the current fsb with core2 not what is going to happen in the future." and I was replying to this: "you have to think the fsb is doing a good job, or the better architecture would be bottlenecked, effectively negating any potential performance gains. "

"To all that might have thought I was attacking them I'm sorry.." and I just wrote out the theory of how the fsb might get bottlenecked. Was not trying to get into a cat fight with anyone either.
March 13, 2008 7:17:15 PM

Dude you don't have to insult me...And if you in a cat fight that's your problem not mine.

If I misunderstood you then my bad, but deff. not a spoon, what ever that means.
March 14, 2008 7:27:12 AM

MU_Engineer said:
Probably not, at least from what most of the tests I've seen run here and other places have said. Look at specFP_rate for 1 thread -> 2 thread scaling as that hits the memory I/O about as hard as anything else I've seen. The FSB only starts to be an issue with synthetic benches such as specFP_rate when running an MCM quad-core CPU on an FSB (such as the Q6600) and only starts to be much of an issue with real programs when running two quad-core MCM CPUs on a single system (such as the E53xx series.)

I'd put your bogging down under running many apps as being the fault of the OS (particularly if you're running Windows XP), only having two cores and >2 intensive threads, amount of RAM, HDD I/O and network I/O ahead of FSB congestion on your E6750.

I hate spec. It's the big lie.
The latest ver maxes out at ~ 25% core usage, but it's suposed to be a processor benchmark?
Intel has it's L1 cache miss rate pegged" as low as 8%" due to smartcache. Most apps run @ 12 to 20%.
Core2's L1 cache miss rate on spec is 2.74% That has got to be the biggest cheat in any benchmark.
Core2's fsb does not bottleneck in spec because the #&*$%^*@ software lines all it's ducks up, to hide any bottleneck.
!