Friday, February 3rd 2012

AMD Slips Out Trinity ULV 3DMark Performance
In a footnote of a slide detailing AMD's Trinity A6 APU for Ultrathin notebooks at the company's Financial Analyst Day event, the new chip's 3DMark performance was revealed. The company was talking about the 17W ULV (ultra-low voltage) variant of the "Trinity" APU in the slide, that's designed for compact notebooks. The 3DMark Vantage performance of the APU was measured to be 2,355 points, in the same test, an Intel Core i5-2537M ULV 17W "Sandy Bridge" processor scored 1,158 points. The AMD chip, hence, emerged with a 103% graphics performance lead.
The slide notes that with an assumed performance increase of 30% by the upcoming "Ivy Bridge" architecture, its 3DMark performance is projected to be 1,505 points. The 17W Trinity chip would still end up with a 56% performance lead. Moving on, AMD even revealed the performance of the high-performance A10 "Trinity" APU with 25W TDP, designed for slightly thicker notebooks. This chip scored 3,600 points in 3DMark, which would effectively make it 136% faster than Ivy Bridge at graphics.
As for CPU performance, it's noted that Intel will clearly have an edge with performance per core, and the upper hand with single-threaded applications, while Trinity could be competitive with multi-threaded applications, as its two-module/four-core APUs will be competitively priced to Intel's two-core/four-thread(HTT) ones. AMD has pulled the presentation off from the public page of AMD-FAD.
Source:
VR-Zone
The slide notes that with an assumed performance increase of 30% by the upcoming "Ivy Bridge" architecture, its 3DMark performance is projected to be 1,505 points. The 17W Trinity chip would still end up with a 56% performance lead. Moving on, AMD even revealed the performance of the high-performance A10 "Trinity" APU with 25W TDP, designed for slightly thicker notebooks. This chip scored 3,600 points in 3DMark, which would effectively make it 136% faster than Ivy Bridge at graphics.
As for CPU performance, it's noted that Intel will clearly have an edge with performance per core, and the upper hand with single-threaded applications, while Trinity could be competitive with multi-threaded applications, as its two-module/four-core APUs will be competitively priced to Intel's two-core/four-thread(HTT) ones. AMD has pulled the presentation off from the public page of AMD-FAD.
107 Comments on AMD Slips Out Trinity ULV 3DMark Performance
My old Sony Vaio VGN-FW45's plays Mass Effect 2 at 720p and max settings.
PS; Sony Vaio VGN-FW45's Intel Core 2 Duo P8700 has 25 watts TDP.
My TI-83 could play Mass Effect 2 at 720p and max settings...(and so can the Sandy Bridge iGPU FYI).
Try something a little more demanding, and a little more modern.
Which Sandy Bridge IGP version i.e. ULV(17 watts, 350MHz**), LV, (500MHz**), desktop(850Mhz**)?
**base clock.
Click on my profile's "System Specs" link for my Silverstone SG07 Mini-ITX's LAN box specs i.e. it has Intel Core i5-2500K with active HD 3000 IGP (allocated 512MB shared memory from UEFI).
To enable HD 3000 IGP, I have used ASUS P8H67-I Rev 3.0 motherboard. My tower PC has ASUS P8P67 Rev 3.0 motherboard i.e. I plan to swap the non-K Intel Core i7-2600 and Core i5-2500K later.
Seriously, 5+ year old games engines are not what I call a stress on a GPU or really what I rush out to play on my brand new computer. Which is why I don't buy anything with an iGPU/APU for gaming.
Yeah right. Now this is a different model, but the A6 have simliar Vantage performance. And you don't do serious gaming on a small notebook anyway, but with these puppies you can do some light gaming even with modern titles.
Nothing hardware decodes Hi10p yet. I do know that for sure. Even if the software could do it, supposedly the current crop of hardware isn't capable anyway. Not sure about all that, but all that matters is that it just doesn't work.
www.clubbleach.org/forums/showthread.php?96051-How-to-Play-10-bit-h264-%28Hi10P%29-video-files
So again, for someone like me, the graphics performance is irrelevent. How many time do you need to be told that I am a fan of no brands whatsoever? I am only a fan of the products that give me the most of what I want for my money. I want cpu power. I don't care who gives it to me.
Keep your fanboy claims to yourself.
Atm, Intel Sandybridge and Ivybridge doesn't support FMA3. AMD Piledriver supports FMA3 and FMA4 instructions.
As for encoding videos on pure CPU, www.tomshardware.com/reviews/fx-8150-zambezi-bulldozer-990fx,3043-17.html
This is using the *flawed* AMD Bulldozer. Atm, the old AMD Bulldozer has issues with the single threaded applications, while in multi-threaded applications it seems to be competitive.
And I'm not exactly sure what point you are trying to make by showing that the Intel is faster in encoding thread for thread at a lower clock speed.
And hi10p is a definite issue for me, thus the entire reason I have brought it up in the first place. It's an issue for anyone that follows anime subbing groups.
From m.hardocp.com/article/2011/10/11/amd_bulldozer_fx8150_desktop_performance_review/7
DVD movie to the iPhone4
The stock 8120 slots between i7-2600K and i5-2500K. I don't see the sound logic to reduce the video player userbase.
I use handbrake, mediacoder and RipBot depending on what I'm trying to accomplish. All of them use official x.264 builds in their code, so that's all I'm interested in when it comes to encoding performance. Most commercial encoders are unoptimized piles of crap.
You are failing to make a point here. This entire time I've been talking about my needs in this class of notebook, and the needs of users like me. GPU performance serves us no purpose, and you just made me aware that the AMD chips are slower thread for thread in multithreaded apps than a hyperthreading Intel with half the number of real cores. I didn't know that. I thought that maybe the 4 core AMDs could take on a 2c/4t Intel. My mistake.
So, barring an extreme price or battery advantage with the AMD laptop, why would I even consider something other than the Intel setup? Hi10p provides up to 25% better compression with no quality loss. More space for more shows. I don't see the sound logic in using an inferior compression algorithm just to gain gpu acceleration when the cpu is already powerful enough to handle it.
Microsoft's latest Bulldozer hotfix treats AMD's Bulldozer module as 1 physical CPU count with 2 logical CPU threads aka hyperthreaded enabled CPU. It's your decision. I already have a 45 watts CPU heavy vs 26 watts GPU gaming laptop and I don't plan on buying another CPU heavy laptop.