SB is here. When will Bulldozer be released?

[tokie]

I decided I don't have confidence in AMD and pulled the trigger on Sandy Bridge.

As Idontcare pointed out, Intel has been sandbagging it in the frequency department. Even if Bulldozer has a slight IPC advantage, Intel can likely still increase frequency to compensate.

 

[Elixer]

++

That is disappointing. I really wish they get new motherboards out there in the meantime.

Yeah, I agree.
They also said there is no shortages of parts for anything, I assume this covers the new 9xx chipset line as well.

It is a shame that nobody asked them what the holdup is.

 

[StinkyPinky]

Regarding the AM3+ boards.

It is also possible they are wanting to release then at the same type as BD to help generate sales (hype etc). Plus perhaps they will come with LightPeak.....(although that's an intel/apple tech so I wonder how that works...)

 

[JFAMD]

On their conference call, they said bulldozer is "on track" for early summer... that means May/June release most likely.

Too bad they didn't say spring,

No change in schedules, everything looks the same.

 

[pitz]

The Linux support for Sandy Bridge, especially the onboard video, apparently, is less than inspiring. If the Sandy Bridge video had good support, it would be an unbeatable platform for Linux.

Anyone think that AMD will roll something out that has the same level of maturity driver-wise, as the current ATI boards under Linux (ie: actually pretty good these days!)?

If so, sign me up. Even if the Bulldozer ends up being slower than the Sandy Bridge chips.

 

[drizek]

No change in schedules, everything looks the same.

So what exactly is the schedule? Can you say whether the motherboards will be released early?

Some questions:

How will the IGP/APU work out? Will AM3+ chipsets have integrated graphics? What about Llano? Is that going to be on AM3+? What kind of upgrade/sidegrade paths will there be for Bulldozer/LLano buyers and Phenom II owners?

Edit:
I guess I should have searched first:

- 990FX:supports two full speed PCI-E 2.0 x16 slot and can be split into four PCI-E 2.0 x8 slot half speedcan be set up to dual quad CrossFireX, the other to support a PCI-E 2.0 x4 slots Six PCI-E 2.0 x1 slots.

- 990X:to support a PCI-E 2.0 x16 slot and can be split into two full speed PCI-E 2.0 x8 slot half speedcan be set up dual CrossFire, also supports six PCI-E 2.0 x1 slots.

- 970:support for a PCI-E 2.0 x16 slot at full speed and can not be split, that is, do not support CrossFire.

South Bridge are SB950, SB920 two models, the main difference is that RAID technology, but are not native USB 3.0.

- SB950: supports four PCI-E 2.0 x1 slots, PCI bus, 14 USB 2.0 interface, six SATA 6Gbps interface, supportRAID 0/1/5/10,can be used with 990FX, 990X, 970.

- SB920: supports four PCI-E 2.0 x1 slots, PCI bus, 14 USB 2.0 interface, six SATA 6Gbps interface, supportRAID 0/1/10,mainly with 970.

Doesn't say anything about Llano, and there doesn't appear to be one with an IGP. It is disappointing to see no USB 3, no PCIe 3.0 and no equivalent to the Z68 HDD/SSD setup.

 

[drizek]

The Linux support for Sandy Bridge, especially the onboard video, apparently, is less than inspiring. If the Sandy Bridge video had good support, it would be an unbeatable platform for Linux.

Anyone think that AMD will roll something out that has the same level of maturity driver-wise, as the current ATI boards under Linux (ie: actually pretty good these days!)?

If so, sign me up. Even if the Bulldozer ends up being slower than the Sandy Bridge chips.

It depends on what you mean by "linux". For general desktop use, Sandy Bridge would be great. If, however, you are referring to compile times for non-binary distros (ie. gentoo), then I think you will want as many threads and cores as you can get your hands on, and Intel made the fairly disappointing move of disabling HT on their affordable SB chips. Honestly though, linux is so efficient that I think any of these chips would be overkill.

 

[pitz]

It depends on what you mean by "linux". For general desktop use, Sandy Bridge would be great. If, however, you are referring to compile times for non-binary distros (ie. gentoo), then I think you will want as many threads and cores as you can get your hands on, and Intel made the fairly disappointing move of disabling HT on their affordable SB chips. Honestly though, linux is so efficient that I think any of these chips would be overkill.

I mean video drivers. Agree with you on the other point, and that is, its an extremely fast CPU that's just fine for everything that can be thrown at it. Hopefully the video drivers will mature very quickly.

I don't even see any Linux drivers at the Intel website for Sandy Bridge's video. Why not?

 

[Martimus]

I decided I don't have confidence in AMD and pulled the trigger on Sandy Bridge.

As Idontcare pointed out, Intel has been sandbagging it in the frequency department. Even if Bulldozer has a slight IPC advantage, Intel can likely still increase frequency to compensate.

That doesn't help you now though. You already bought the processor, so if intel increases clock speed, you are SOL.

 

[Elixer]

So what exactly is the schedule? Can you say whether the motherboards will be released early?
.

A simple yes or no will be nice, but I don't think JF has those answers, he is on the server side, not desktop.

Regarding the AM3+ boards.

It is also possible they are wanting to release then at the same type as BD to help generate sales (hype etc). Plus perhaps they will come with LightPeak.....(although that's an intel/apple tech so I wonder how that works...)

Highly doubtful, the hype will still be there, it is just that releasing AM3+ boards now generates more sales for mobo partners, and also possibly can generate more CPU sales for people waiting for BD.

I still think this is a win-win for everyone, but nobody is saying anything about a possible release of AM3+ before BD is launched.

 

[Ajay]

And therefore, Bulldozer must compete very well.

So the same as Gulftown, but not necessarily better. I guess we'll have to see what the performance/watt is (for those that care).

 

[JFAMD]

I know the schedules but Q2 for desktop launch, Q2 for start of server production and Q3 for server launch is all I can say.

 

[bryanW1995]

Unfortunately for AMD, this "plan B" has been available to Intel since C2D was launched.

They consistently sandbagged the clockspeed potential of their chips while maximizing yields and gross margins because AMD has yet to give them "a run for their money" since summer 2006.

And that is the sad reality for AMD...trumping today's SandyBridge SKU's with Bulldozer won't be enough because today's SB SKU's are all underclocked, immensely so.

A BD that thumps SB will simply mean lower margins for Intel as they will boost clockspeeds at each given pricepoint, thumb BD in a return salvo of gunfire, and take the yield (binning) hit.

As customers we only stand to win, but AMD is dealing with constantly shifting goalposts as they try and score a touchdown. That sucks for them, will be hard to earn a sense of accomplishment if you are on team BD.

maybe pushing into other markets isn't such a bad idea for amd after all...

 

[bryanW1995]

That doesn't help you now though. You already bought the processor, so if intel increases clock speed, you are SOL.

hopefully he got a K series cpu and a p67 mobo. based upon past experience, SB respins are unlikely to up the overall top end more than 200-300 mhz anyway, so he'd be quite close the top already. I think that IDC's point is that even if BD kicks SB's tail at stock speeds, intel can just roll out an EE chip at 4.5 or higher if necessary and cascade their cpus down from there. Most likely scenario imho is that BD gets into the $300 market but amd shies away from the $500-$1000 market b/c they know what intel's response would be, anyway, and it wouldn't make anybody more money. also, they could really piss of intel if they tried for that high end market and intel could go after the $100-$200 market that amd has been quite competitive in lately. how hard could it be for intel to tell its board partners how to hack in an OC for non-k chips for example?

 

[grimpr]

The Linux support for Sandy Bridge, especially the onboard video, apparently, is less than inspiring. If the Sandy Bridge video had good support, it would be an unbeatable platform for Linux.

Anyone think that AMD will roll something out that has the same level of maturity driver-wise, as the current ATI boards under Linux (ie: actually pretty good these days!)?

If so, sign me up. Even if the Bulldozer ends up being slower than the Sandy Bridge chips.

Take a look at Phoronix findings through the land of Linux graphics.

http://www.phoronix.com/vr.php?view=15638
http://www.phoronix.com/vr.php?view=15614
http://www.phoronix.com/vr.php?view=15604
http://www.phoronix.com/vr.php?view=15592
http://www.phoronix.com/vr.php?view=15574

 

[Ajay]

hopefully he got a K series cpu and a p67 mobo. based upon past experience, SB respins are unlikely to up the overall top end more than 200-300 mhz anyway, so he'd be quite close the top already. I think that IDC's point is that even if BD kicks SB's tail at stock speeds, intel can just roll out an EE chip at 4.5 or higher if necessary and cascade their cpus down from there. Most likely scenario imho is that BD gets into the $300 market but amd shies away from the $500-$1000 market b/c they know what intel's response would be, anyway, and it wouldn't make anybody more money. also, they could really piss of intel if they tried for that high end market and intel could go after the $100-$200 market that amd has been quite competitive in lately. how hard could it be for intel to tell its board partners how to hack in an OC for non-k chips for example?

I hope AMD can get an FX version of BD in @ $500-$700, mainly so that they can make a decent profit and afford the R&D to stay competitive w/Intel. That might not start a price war since Intel's high end offerings are priced higher than that.