Tuesday, April 10th 2012

Trinity (Piledriver) Integer/FP Performance Higher Than Bulldozer, Clock-for-Clock

AMD's upcoming "Trinity" family of desktop and mobile accelerated processing units (APUs) will use up to four x86-64 cores based on the company's newest CPU architecture, codenamed "Piledriver". AMD conservatively estimated performance/clock improvements over current-generation "Bulldozer" architecture, with Piledriver. Citavia put next-generation A10-5800K, and A8-4500M "Trinity" desktop and notebook APUs, and pitted them against several currently-launched processors, from both AMD and Intel.

It found integer and floating-point performance increases clock-for-clock, against Bulldozer-based FX-8150. The benchmark is not multi-threaded, and hence gives us a fair idea of the per core performance. On a rather disturbing note, the performance-per-GHz figures of Piledriver are trailing far behind K12 architecture (Llano, A8-3850), let alone competitive architectures from Intel.
Source: Expreview
Add your own comment

115 Comments on Trinity (Piledriver) Integer/FP Performance Higher Than Bulldozer, Clock-for-Clock

#26
eidairaman1
The Exiled Airman
your post here is offtopic
Posted on Reply
#27
Tatty_Two
Gone Fishing
OneMoarthat is a little late
personally at this stage I think we(the consumer) need to make it VERY CLEAR to AMD that son is disappoint and they need to step it up or step out
AMD kept a fairly tight lid on bulldozer probably because they knew they where not gonna be able to keep up with there hype and it backfired ...
when a manufacturer doesn't talk to you about there product's in development it means one of two things
either they are making some really great and don't want to let the competition in on it
OR more often then not they are having trouble and are hoping you wont notice
the problem here is the majority of AMD's consumer base happen to know what the fuck they are talking about and wont have the wool pulled over there eyes very easily
I agree with much of what you have said here, however surprisingly AMD's market share since Bulldozer was released has actually increased (Q4 2011), more so in the server market (although full year figures show a slight decrease in Intel's Market share with a slight increase to AMD's.... across the whole form factors), albeit not by much interestingly though, intels dominance was reduced by the same amount as AMD's gain, suggesting at least that AMD is managing to appeal to some OEM's, if only a minority.
Posted on Reply
#28
Dent1
IMO this article is fail.

Trying to compare a budget laptop quadcore Trinity APU to a high end desktop octocore Bulldozer is utter fail.

Until I see a desktop Piledriver, without any gimps with full cache I'm not passing any judgement.

Edit:

Also, its great that a low end Trinity is spanking the Bulldozer, which is promising for the full blown desktop Piledriver.
Posted on Reply
#29
xenocide
Dent1IMO this article is fail.

Trying to compare a budget laptop quadcore Trinity APU to a high end desktop octocore Bulldozer is utter fail.

Until I see a desktop Piledriver, without any gimps with full cache I'm not passing any judgement.

Edit:

Also, its great that a low end Trinity is spanking the Bulldozer, which is promising for the full blown desktop Piledriver.
It's more or less to allow a frame of reference. Trinity APU's are going to use the same Piledriver cores as the FX line, just without the L3 Cache, and minor tweaks, as I understand it. So assuming you can get a nice 5% bump from the tweaks, you can say that if the PD cores in APU's are 10-20% faster than BD cores, then the PD Cores in desktops should be 15-25% faster than the BD cores in desktops.

It's not so much comparing Laptop APU's to Desktop CPU's, as it is comparing what's under the hood and what it could mean for Desktop CPU's in the future from AMD.
Posted on Reply
#30
eidairaman1
The Exiled Airman
xenocideIt's more or less to allow a frame of reference. Trinity APU's are going to use the same Piledriver cores as the FX line, just without the L3 Cache, and minor tweaks, as I understand it. So assuming you can get a nice 5% bump from the tweaks, you can say that if the PD cores in APU's are 10-20% faster than BD cores, then the PD Cores in desktops should be 15-25% faster than the BD cores in desktops.

It's not so much comparing Laptop APU's to Desktop CPU's, as it is comparing what's under the hood and what it could mean for Desktop CPU's in the future from AMD.
Your post makes sense! :toast:

15-25% is definitely a better than nothing gain considering BD-FX patching for win 7 didnt do anything. BD-FX being a server chip under the hood. Piledriver and Trinity may verywell be totally aimed at the desktop user and not a workstation/server environment like BD-FX and Opteron 3200,4200,6200 are. (Despite 3200 apparently being faster than BD-FX)
Posted on Reply
#31
joyman
The improvements look ok, I just hope the price will be improved also in the right direction :) So finally I can replace my 960T with some nice Piledriver chip.
Posted on Reply
#32
Hustler
xenocidethen the PD Cores in desktops should be 15-25% faster than the BD cores in desktops.
So what...that will just bring them up to Phenom II levels of performance (maybe slightly faster)...still an utter fail.
Posted on Reply
#33
ensabrenoir
Anyfhing bd related = flame war ...those words should not be used together . Hoping pd is actually better...pipe dream at most because itll make intel release something even more awesome. I figure they have several aces they havent played because...well u know ..no reason to no comp. Anyway carry on with the other worldly performance increase fantasies.
Posted on Reply
#34
xenocide
HustlerSo what...that will just bring them up to Phenom II levels of performance (maybe slightly faster)...still an utter fail.
Phenom II level per thread performance, with substantially better overclocking, a greatly improved IMC, and support for tons of new Instruction Sets makes for a pretty solid chip, especially if the price stays low and they can get the power consumption under control. It means the AMD FX line is no longer 13 junk CPU's and the FX-8120 which is a steal.
Posted on Reply
#35
Fourstaff
Still waiting for AMD to improve performance per thread and power consumption, without these two Intel will almost certainly get my custom instead.
Posted on Reply
#36
Kärlekstrollet
A high IPC isn't always that great, a good IPC that's energy effective and scales well with overclocking.
In the end does Trinity outperform Llano and draw less power, mission well done AMD.
Posted on Reply
#37
Melvis
seronx:shadedshu , It's like you don't even know how FX eight cores even score... :laugh:
:confused: Ummm i think i do, and the 50 benchmarks around on the web show it losses 95% against the 2600k so i don't know what ya trying to say here sorry?

And if ya ask have i ever played with Bulldozer then the answer is yes a 8120, so i know how they work and perform.

Bring it up to par with what the 2600K is in most benchmarks/games etc then id be happy, thats all im asking for, is that to much?
Posted on Reply
#38
zenlaserman
seronxI like how everyone lies... :toast:
I like how you constantly delude yourself.

What I don't like is how you try to spread it like a wet gremlin. :eek:
Posted on Reply
#39
Atom_Anti
Guys, good balanced CPU / GPU performance is more important than just a very fast CPU;).
Posted on Reply
#40
Fourstaff
Atom_AntiGuys, good balanced CPU / GPU performance is more important than just a very fast CPU;).
Which is why you get a fast cpu coupled with a fast gpu rather than caring about APUs ;)
Posted on Reply
#41
wiak
well the reason trinity and liano uses more power is because they have real graphics, that can be compared to a mid range radeon hd
Posted on Reply
#42
xenocide
wiakwell the reason trinity and liano uses more power is because they have real graphics, that can be compared to a mid range radeon hd
Llano didn't have any "mid-range" level iGPU's. If I recall the best iGPU was on par for about an HD6450 or under best circumstances a 5570. That's entry-level at best. The Trinity APU's have a 7660D, which means it should be substantially better, but still nothing compared to what most people classify as "mid-range".
Posted on Reply
#43
Aquinus
Resident Wat-man
xenocideLlano didn't have any "mid-range" level iGPU's. If I recall the best iGPU was on par for about an HD6450 or under best circumstances a 5570. That's entry-level at best. The Trinity APU's have a 7660D, which means it should be substantially better, but still nothing compared to what most people classify as "mid-range".
It's better than an Intel HD 3000 iGPU. :wtf:
Posted on Reply
#44
Fourstaff
AquinusIt's better than an Intel HD 3000 iGPU. :wtf:
That would be a pointless argument, since that we expected Llano to be able to run newer games at decent quality, whereas if we are getting Intel's chip we know that we must get a discrete. I think 3rd gen APUs will be powerful enough to be considered an alternative to discrete, but as of now APUs are only powerful enough to run counter strike and farmville, with Trinity powerful enough to run everything on low.
Posted on Reply
#45
Aquinus
Resident Wat-man
FourstaffThat would be a pointless argument, since that we expected Llano to be able to run newer games at decent quality, whereas if we are getting Intel's chip we know that we must get a discrete. I think 3rd gen APUs will be powerful enough to be considered an alternative to discrete, but as of now APUs are only powerful enough to run counter strike and farmville, with Trinity powerful enough to run everything on low.
...but that is what you pay for when you get a CPU + GPU for 100 USD. For how much you're paying, it's a good bargain, not a high-end solution.
Posted on Reply
#46
alwayssts
OneMoar...more often then not they are having trouble and are hoping you wont notice the problem here is the majority of AMD's consumer base happen to know what the fuck they are talking about and wont have the wool pulled over there eyes very easily
First I was like: Holy fuck, punctuation rape Batman.

Then I was like: Man makes a very valid point.
Posted on Reply
#47
Hustler
xenocidePhenom II level per thread performance, with substantially better overclocking
Well seeing as a Bulldozer has to run at about 4.2Ghz just to match my Phenom II at 3.8Ghz, the fact that the average Bulldozer only goes up to about 4.5Ghz (with non exotic cooling methods) will still only make them about 10% faster, whilst at the same time using loads more power for very little return.

Even with these new refined PD cores @ 4.5Ghz, i would only be looking at about a 25% overall improvement over my Phenom II , whereas a 2500k @ 4.5Ghz would be nearer 40-50% faster than my Phenom II.
Posted on Reply
#48
Vulpesveritas
HustlerWell seeing as a Bulldozer has to run at about 4.2Ghz just to match my Phenom II at 3.8Ghz, the fact that the average Bulldozer only goes up to about 4.5Ghz (with non exotic cooling methods) will still only make them about 10% faster, whilst at the same time using loads more power for very little return.

Even with these new refined PD cores @ 4.5Ghz, i would only be looking at about a 25% overall improvement over my Phenom II , whereas a 2500k @ 4.5Ghz would be nearer 40-50% faster than my Phenom II.
And we don't know how high they can go yet. And on a bright side, this may mean AMD sees the BD architecture as having higher max IPC if it does match PII, seeing as K12 STARS only had what, a 5% IPC boost at 32nm?
Posted on Reply
#49
faramir
HumanSmokePersonally I won't believe a word about about Trinity/Piledriver core performance until I hear it from John Fruehe....oh wait
Joke Fruehe ;)
Posted on Reply
#50
Patriot
xenocideThat table is odd, why are several CPU's listed twice when thers are only listed once?
So all are listed at turbo...maybe they did two runs?

see if you get anything more from the source... either way performance is nice for APU...especially without L3...

translate.google.com/translate?hl=en&sl=zh-CN&u=http://www.expreview.com/19073.html&ei=gT2ET53xMIeg9QTSv4nUCA&sa=X&oi=translate&ct=result&resnum=1&ved=0CCYQ7gEwAA&prev=/search%3Fq%3Dhttp://www.expreview.com/19073.html%26hl%3Den%26safe%3Dactive%26client%3Dfirefox-a%26hs%3DO8H%26rls%3Dorg.mozilla:en-US:official%26prmd%3Dimvns
Posted on Reply
Add your own comment
May 5th, 2024 15:03 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts