Just a little curious, does it make unbearable noise when you run bitcoin on both?

Are you able to adjust voltage for the discrete?

A laptop with Llano APU and 400 shaders.
With additional gpu and 480 shader.

Will I be able to mine with all 880 shaders?

No guessing, I need to know for sure.

I'm very sorry but now you're making yet another mistake.

Do you know what the HD 6550M is? It's a rebrand of the HD 5570. Codename is Redwood, UVD is 2.2.
It's from the HD 5xxx family, so it's no surprise that it was launched so fast.



It's nothing personal, but I've just noticed that you are posting many, many errors.

I have probably become too autistic  for my own good.

Do I need to? Read what Jack actually said.

"Also, AMD and Nvidia account for nearly all of TSMC's business."
This is so blatantly wrong, it's like claiming the earth is flat.
If you don't understand where I'm coming from, Jack made many bullshit statements before, like claiming the HD6990 is a 5 gigaflop card. This one is really bad misinformation.
Or when he claimed Barts was VLIW4
Or like when he claimed that mobiles parts are the ones released first
All of that is pure misinformation, especially the first one, and just now he wanted us to believe AMD and nvidia are TSMC's only customers...  



Okay so now you're backtracking, I'll accept that. Before, you claimed AMD and Nvidia accounted for all of TSMC's business, suddenly now you were only talking about the 28 nm processes all along.

Here's where you went wrong. High end videocard chips are very low volume, and TSMC has many other customers, so a low number like 1% could even be natural. But you didn't know that.
So you erroneously assumed that 1% was "proof" that Nvida and AMD cancelled their orders. This was the big fallacy on your part all along.

You posted a link that stated "28nm chips will only account for some 1% of their revenues" then stated "AMD/Nvidia have canceled their major orders". A feather becomes a chicken.


I was saying that 1% could still mean a lot of wafers. The reason for this is that TSMC has a lot of other customers who utilize their older 55 and 65 nm processes.
Then you attempted to argue against this, by claiming that AMD and Nvidia were the only customers of TSMC anyway in order to give the misleading picture of that 1 percentage amounting to nothing. Which is blatantly wrong. They have many other customer, Which is EXACTLY the reason why that 1 percentage could still be a lot of wafers.



TSMC's projected 28 nm revenues going from 2.5% to 1% does not mean the world is going to end.

This is not proof that AMD has cancelled their orders.

Okay. You don't know what you're talking about. It's best to just step back instead of spreading more misinformation.

Your world may revolve around video cards, but the real world does not do that.



Southern Islands taped out already back in February...

Going from 2.5% to 1%.
How can you be sure that 1 lost percentage didn't belong to nvidia?
As I said, everybody already knows nvidia sucks at new manufacturing processes.

No, that was the 5850.

5830 has higher clocked shaders and use more power. They should be around 180 watt if I recall correctly.

Where does it say this?  


1% could still be a lot of wafers. Do you have any data of how many of their existing customers are still utilizing their 65 nm process for example?
It was never a big secret that high end videocards have always been low volume.


How can you be sure this has any relevance to AMD at all? Southern Islands taped out already back in February...

The cpu speed is irrelevant. What you look at is how many shaders there are, and how fast the shaders run.

400 shaders at 600 mhz is the strongest Llano you can get.


Best to deactivate the cpu mining and only mine with the graphics shaders.
Overclocking the shaders from 600 Mhz should boost your hash rate far more than the cpus ever could achieve.


It's vliw5 at 32 nm soi. It's far more power efficient than any other graphics card in existence currently.

To truly test cpu, use these flags:
-g no -t 8

It will tell Ufasoft to use 8 threads, and to not use graphics cards.




Agreed.

A few months back, these would be insanely valuable. But with the current difficulty... I'm kind of doubting they are worth the extra cost.

She probably got annoyed by the noise or the electricity bill. 

Bios editing. It's pretty safe to do when you have a bios switch.

My second favourite would be Ati Tray Tools, but it doesn't support multiple cards.

Not necessarily so, what you fear is completely unfounded. Traditionally, ATI/AMD has always been better than nvidia at utilizing newer manufacturing processes.

HD 4770 was built on 40 nm. It was released in April 2009.
GT 240 was built on 40 nm. It was released in November 2009.
ATI was 7 months ahead.

HD 5870 was released September 2009.
Gtx 480 was released March 2010.
ATI was 6 months ahead.

HD 4870 was 55 nm, released June 2008.
Gtx 285 was 55 nm, released January 2009.
ATI was 7 months ahead.

Do you see a pattern? The two companies specializes in different competencies. It would be unwise to attribute some of nvidia's weaknesses to the ATI team.

AMD had a momentarily weakness when TSMC cancelled the 32 nm process. This incident screwed ATI over because their strength was taken from them and they were trapped on 40 nm.
But this time, it's back to business as usual.


The SI tapeout happened in February

Like I said, ATI tends to be more competent on newer processes, so nvidia's problems may not necessarily be ATI's problems.

I'm fairly certain they will. And if the PSU doesn't have enough pci-e power plugs, the graphics cards always have power plug converters.

I tend to agree. As long as you only have 1 graphics card.

I never managed to make Ati Tray Tools work for multiple cards.

Mobility chips have traditionally always been slower to release. It goes like this:
Desktop cards first
Mobile cards second
Workstation cards last


That's not what m3sSh3aD meant. He's guessing that they will release the entire top to bottom range in Q3/Q4 (using slightly inferior manufacturing process, but available earlier).

There are some HD 5450 out there on PCI. But they aren't worth it.

I know. But AMD was locked at the 40 nm process at that time.
Barts was 1120 shaders.
Turks was 480 shaders.
They needed something inbetween. Instead of redesigning a new 800 shader asic at 40 nm, they reused their current chip. Redesigning new chips at the same shader numbers would waste millions of dollars.


This time it's different, they are no longer locked at 40 nm, and you know that.

I undervolt my stuff, so nothing has died yet.

It's still up in the air whether the new generation from AMD will be introduced in 3rd quarter or 4th quarter. It could be either at this point.


Even if the VLIW4 microarchitecture may be reused, it won't be fair to call it rebadging. You can only call it that if they take the 40 nm Cayman, and let it remain 40 nm, yet call it 78x0 or something like that.

But this time, hopefully everything will be shrunk by two halfnodes down to 28 nm, in this case they will be entirely new chips.