HP Envy Sleekbok sneak premieres with AMD Trinity

[IntelUser2000]

Was looking at the same benchmarks for Bulldozer, for comparison an FX-8150 gets 31678 in CPU Queen according to the AnandTech review. With half the cores and 56% of the clock speed (2.0GHz vs 3.6GHz) it should get about 8870. A10-4665M gets 11547, so branch prediction is 30% better? Looks like there might be some pretty solid improvements over BD at least.

Scaling isn't linear at all with number of cores. For example for the Phenom II:

Phenom II X4 975 BE 3.6GHz: 26351
Phenom II X6 1100T: 3.3GHz: 32465

That's only 34% gain for 50% more cores after normalizing for frequency difference.

Really ? These are harvested bargain-bin Trinity parts considering what we should be having at 17W and 35W and even these will only be available in June ?

Let me guess:

April-Shipment starts
May 15-Announcement/NDA lifts
June 20-Product availability

 

[Joseph F]

To kill every tenth man, by random lot? Sure about that word?

That's what I was thinking, as well. Us pedants... 😀

 

[ShadowVVL]

wonder if the desktop apus will be available at same time or earlier.

not sure if it matters they seem to be going nowhere fast.

 

[frostedflakes]

Scaling isn't linear at all with number of cores. For example for the Phenom II:

Phenom II X4 975 BE 3.6GHz: 26351
Phenom II X6 1100T: 3.3GHz: 32465

That's only 34% gain for 50% more cores after normalizing for frequency difference.

Let me guess:

April-Shipment starts
May 15-Announcement/NDA lifts
June 20-Product availability

Thanks, figured I was making an incorrect assumption somewhere. 😀

 

[Gideon]

wonder if the desktop apus will be available at same time or earlier.

not sure if it matters they seem to be going nowhere fast.

Rumors are the desktop variants and corresponding FM2 motherboards will be available later, somewhere in Q3.

 

[Olikan]

it's from an amd blog, take it with a grain of salt

Testing performed by AMD Performance Labs as of 2/17/2012 using the 2012 “Manaslu” Reference Design with 2 x 2G DDR3 1600 MHz memory, 250GB 5400rpm HDD (SATA) and Windows 7 Home Premium. Results from 3D Mark11 Performance show the AMD A6-4455M APU to score 626 3D marks. The high-end Intel Core i5 Ivy Bridge mobile processor is projected to show a 3D Mark11 Performance score of 530. This assumes approximately the same gain in graphics performance from Nehalem to Sandy Bridge and is extrapolated based on Intel’s public statements of anticipated performance gains. This provides the AMD A6-4455M APU an 18.1% performance advantage over the projected Intel Ivory Bridge score. The 3D Mark 2011 Performance score for the A10-4600M APU is 899. This is 70% better than the projected Ivy Bridge score.

http://blogs.amd.com/play/2012/05/10/meet-the-2nd-generation-amd-a-series-apu-codenamed-%E2%80%9Ctrinity%E2%80%9D-powering-your-ultrathin-and-light-gaming-machine/

actually, what's the true score of ivy in 3D Mark 2011 ?

 

[IntelUser2000]

There's no public results out yet for the 17W and dual core Ivy Bridge. The quad core version gets 700 something. The 17W version is likely to end up lower, but its not known by how much.

I don't know how they extrapolated it. Neither Arrandale nor Sandy Bridge can run 3DMark11 so its hard to compare from that.

 

[Abwx]

According to AMD , a Trinity A6 ULV 17W has a 2355 3Dmark vantage score ,
compared to 1158 for a 17W i5 ULV 2537M.

With an assumed 30% increase , as per Intel s claim ,IB is projected
to score 1505 , hence Trinity is projected to have 56% advantage
over the latter.

Score for a 25W Trinity A10 LV is 3600.

 

[polyzp]

This looks like great news to me!

Look at the results carefully

IPC of trinity BEATS IPC of LLano (K 10.5) because both the A10-4655m AND the A6-3400m turbo to the exact same 2.3 ghz clock. This just implies the IPC jump from current bulldozer will be huge! If the math is correct, and IPC is roughly equivilent to K10.5, this is good news! If the FX 8350 even vaguely matches the IPC of K10.5, this would be fantastic! (Given FX 8150 falls behind phenom II IPC by ~20-25%)

 

[Abwx]

IPC is still a big unknown despite a few leaks since Trinity s actual
frequencies are not known for sure.

Seems that FP perfs will be almost unchanged compared to BD cores
while Integer throughput will see more tangible gains.

Comparison with Llano is known to be favourable , since you already have the info.....

http://amdfx.blogspot.fr/2012/04/mobile-trinity-coming-may-15th-early.html