AMD shows off BD architecture at Hotchips

[Chinoman]

But wouldn't it be the Enthusiasts who'd be willing to spend the money for similar priced Chip if the performance was close to Intels? I'd think the average consumer would rather go with an Intel if the price was the same due to brand recognition and Intels proven track record.

I think incidents like the Sandy Bridge chipset bug showed Intel isn't impervious to bad luck.

 

[ed29a]

But wouldn't it be the Enthusiasts who'd be willing to spend the money for similar priced Chip if the performance was close to Intels?

Yes ... and? The money Intel/AMD make off enthusiasts is a tiny drop in a bucket compared to big OEMs like Dell or HP. Enthusiasts, while they can nerdrage all day long, at the end of the day this nerdrage has very little or no effect on AMD or Intel.

 

[Piano Man]

But wouldn't it be the Enthusiasts who'd be willing to spend the money for similar priced Chip if the performance was close to Intels? I'd think the average consumer would rather go with an Intel if the price was the same due to brand recognition and Intels proven track record.

Totally agree. I would love for BD to be like K7, which is when I built a non-Intel machien since it was a no brainer. But if its only close to Intel, with the same price point, I'll probably stick to SB/SB-E for my next build.

 

[SonicIce]

bd suks /thread

 

[Red Dawn]

Yes ... and? The money Intel/AMD make off enthusiasts is a tiny drop in a bucket compared to big OEMs like Dell or HP. Enthusiasts, while they can nerdrage all day long, at the end of the day this nerdrage has very little or no effect on AMD or Intel.

Lol @ nerdrage

I guess they can push the 8/6 core aspect of the chip to the consumers

 

[Red Dawn]

I think incidents like the Sandy Bridge chipset bug showed Intel isn't impervious to bad luck.

Which probably was only a big issue to the Enthusiast crowd who are usually the first adopters.

 

[grkM3]

Do you have any numbers and/or statistics to back up your claim or you are pulling stuff out of your rear?

Ill look for the article I read stating how well sandys are selling even with there recall they had at launch.

A lot of laptop makers are also waiting for there ivys to put out there ultra slim lines that get better battery life than the chips out now.

 

[bryanW1995]

AMD was just as equally hush-hush about phenom II. I even remember Anand criticize AMD for being so quiet and how poor the lunch was. AMD does not do well promoting new products period.

I don't think that AMD's hush-hush attitude is necessarily indicative of poor performance. However, the enormous (and continuing) launch delay is a huge tell. As many others have stated, intel probably has a very good idea about performance, anyway, so what's the use in leaking certain benchmarks if they will put AMD in a positive light? And more importantly, how much of an improvement would BD have to be over 2600k to keep intel from responding with a higher-clocked model? I've postulated that intel could easily release a 4.2 ghz 2700/2800/2900k cherry-picked cpu, but in reality that could probably go even higher than that if necessary. I hate to say that the odds are zero, but BD's chances of beating intel this round are about as close to zero as you're gonna get.

 

[ElFenix]

after a 3000 post thread and a 1200 post thread we still have no idea about what this thing will actually do. too bad you can't just get in a hibernation chamber or something. though you might have to deal with otters.

 

[soccerballtux]

Yes ... and? The money Intel/AMD make off enthusiasts is a tiny drop in a bucket compared to big OEMs like Dell or HP. Enthusiasts, while they can nerdrage all day long, at the end of the day this nerdrage has very little or no effect on AMD or Intel.

Lol @ nerdrage

I guess they can push the 8/6 core aspect of the chip to the consumers

pretty sure they make more than that off us judging by how much better AMD was doing profit wise back when Athlon64 was king..

 

[darkswordsman17]

I don't think that AMD's hush-hush attitude is necessarily indicative of poor performance. However, the enormous (and continuing) launch delay is a huge tell. As many others have stated, intel probably has a very good idea about performance, anyway, so what's the use in leaking certain benchmarks if they will put AMD in a positive light? And more importantly, how much of an improvement would BD have to be over 2600k to keep intel from responding with a higher-clocked model? I've postulated that intel could easily release a 4.2 ghz 2700/2800/2900k cherry-picked cpu, but in reality that could probably go even higher than that if necessary. I hate to say that the odds are zero, but BD's chances of beating intel this round are about as close to zero as you're gonna get.

The issue is that I don't see that happening as it would actually likely hurt Intel more than AMD. Intel wants you to buy the higher end platform if you want better performance. Its not impossible, but I'd say its doubtful that if Bulldozer beats Sandy Bridge that Intel will combat it on 1155.

 

[nonameo]

The issue is that I don't see that happening as it would actually likely hurt Intel more than AMD. Intel wants you to buy the higher end platform if you want better performance. Its not impossible, but I'd say its doubtful that if Bulldozer beats Sandy Bridge that Intel will combat it on 1155.

Not only that, even if AMD makes a uber proc that kicks the crap out of even SB-E, they don't have the manufacturing capacity to make enough CPUs to fill demand. So, Intel will fill the remaining orders at well... whatever price they feel like. All while still having a majority of the market.

example: look at all the supply problems AMD is having with llano now and were having with ontario/zacate when it first came out.

 

[3DVagabond]

Not only that, even if AMD makes a uber proc that kicks the crap out of even SB-E, they don't have the manufacturing capacity to make enough CPUs to fill demand. So, Intel will fill the remaining orders at well... whatever price they feel like. All while still having a majority of the market.

example: look at all the supply problems AMD is having with llano now and were having with ontario/zacate when it first came out.

Seems like AMD is constrained by their finances. They're probably being "safely conservative" so as not to contract more manufacturing resources than they can pay for should something go wrong on their end. I would guess that you have to book Fabs ~12mos (or maybe more) ahead of time. If they ended up with nothing to make they still would likely have a Fab bill to pay.

 

[BlueBlazer]

Yes ... and? The money Intel/AMD make off enthusiasts is a tiny drop in a bucket compared to big OEMs like Dell or HP. Enthusiasts, while they can nerdrage all day long, at the end of the day this nerdrage has very little or no effect on AMD or Intel.

Agree. However its the enthusiasts (and retail) where higher end CPUs are likely to be bought. Majority OEM machines (from Dell, HP, etc) often feature lower end CPUs. Usually higher end CPUs ends up in customized gaming PCs (CyberPower, Alienware, etc). AMD long before the big Dell deal, had (if not mistaken) Acer and HP during the K7 days (also that time Apple was using the PowerPC and hasn't made the switch to Intel yet).

pretty sure they make more than that off us judging by how much better AMD was doing profit wise back when Athlon64 was king..

And the prices were higher then. Also take into consideration the prices of Athlon FX and X2. Nowadays AMD does not have such pricing luxury (being constrained to under $200 range).

 

[AtenRa]

I agree there are parallels that can be drawn between the 90nm Athlon X2, 65nm Athlon X2, 65nm Phenom and the 45nm Phenom II X4, 32nm Phenom II X4 (llano), 32nm Bulldozer.

We didn't know to rationalize it as such at the time, but when those initial 65nm Athlon X2's came out and they could not even meet the clockspeeds of their 90nm predecessors it turned out to be a harbinger of the health of 65nm and what that meant for 65nm phenom clocks.

Now look at Llano clockspeeds versus power...what should we be expecting from bulldozer? Miracles that defy device physics? (maybe...prescott sucked at 90nm but Dothan did not)

From the picture below, it seams that Llano's core has a higher transistor density than BD.

Llano's core size @ 9,69mm2 (ext L2) with 35M transistors = 35/9,69 = 3,61M transistors per mm2

Bulldozer core size @ 19,42mm2 (ext L2) with 67M transistors = 67/19,42 = 3,45M transistors per mm2

Since both cores are manufactured with the same process at the same fab, we have two explanations,

BD uses bigger (higher performance, more efficient at higher voltages and frequencies) transistors than Llano or, Llanos core design helps with higher density.
I will vote for the first

For comparison, SB 55/16,5 = 3,33M transistors per mm2

 

[ViRGE]

From the picture below, it seams that Llano's core has a higher transistor density than BD.

Llano's core size @ 9,69mm2 (ext L2) with 35M transistors = 35/9,69 = 3,61M transistors per mm2

Bulldozer core size @ 19,42mm2 (ext L2) with 67M transistors = 67/19,42 = 3,45M transistors per mm2

Since both cores are manufactured with the same process at the same fab, we have two explanations,

BD uses bigger (higher performance, more efficient at higher voltages and frequencies) transistors than Llano or, Llanos core design helps with higher density.
I will vote for the first

For comparison, SB 55/16,5 = 3,33M transistors per mm2

What's the ratio of SRAM and interconnects to logic though? SRAM is more dense than logic, whereas interconnects consume large amounts of space with few transistors.

 

[BlueBlazer]

A bit of news >> AMD Names Rory P. Read President and CEO