• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

First AMD Llano “A8-3800 Pictures and Benchmarks Exposed?

2.4GHz stock seems extremely low for a desktop quad-core. The model name does not match up with the previously leaked models. Is this just an ES?
 
^

Its a GPU and Cpu so it has to stay cool and in OEM power requirments. Let alone it clocks up to 2.7Ghz when needed. Which is not bad for a early first gen.

Plenty to power the built in GPU and do what 99% of people out there need and some. I take it over my "new" HP I have at work anyday.
 
Not bad for a 1st gen APU. Solid gaming IMO, maybe with a bit more tweaking (oc) we can get better numbers from the setup
 
2.4GHz for quad core (Deneb) at 32nm (yes + 400 SPs for the IGP) is too low.


The Diff between 2.4 and 3ghz is very little in any gaming mark. Remember this is not supposed to be a big power house system but a desktop that everybody buys cheaply.

i.e. compare this to a OEM system with onboard video.
 
The Diff between 2.4 and 3ghz is very little in any gaming mark

You dont take the APU only for Gaming

Remember this is not supposed to be a big power house system but a desktop that everybody buys cheaply.

AMD Llano A8 Desktop chips will be 100W and you can already buy a cheap AMD Quad Athlon + HD5570 that will be faster than the 2.4GHz Llano in the link.

i.e. compare this to a OEM system with onboard video.

I was specifically talking about the CPU frequency and not the IGP.
 
The CPU Mark score is 187. Considering a Celeron 420 single core scores more than twice that, I am suspicious of these numbers.

Could be software not supporting the CPU as it has not been updated to recognize it. This is why companies do not want their tech to leak as they know there are some programs that will need to be updated to recognize/support it.

But being this is an OEM type cpu; as long as it works well and is cheap enough OEMs will eat it up. Probable see a lot more HTPC type systems.
 
The CPU Mark score is 187. Considering a Celeron 420 single core scores more than twice that, I am suspicious of these numbers.

Yeah, something's screwy. At 3GHz, all it can pinch out of SuperPi 1M is 25s which is a fair bit off of where it should be. 54s at stock makes me think it's not kicking out of throttled down state.

I hope this is some BIOS bug or something due to it being pre-prod hardware.
 
Posted in the original thread in coolaler..

"Hi all. As I don't speak Chinese ,I will write in English so I hope you understand me .
First of all your Llano is not working at 5.4Ghz,there is a bug in bios.Actual clock speed is around 3Ghz (when you see 5.4GHz in CPUz). The ratio between REAL clock and what you see is around 1.75x. So when you saw 2.4Ghz it was actually locked at 1.4GHz,when you saw 4.7GHz it was performing tests at actual clock of ~2.68Ghz.. You get the picture .

It is evident from your 3dmark06 CPU test,CPUmark99,superpi and ESPECIALLY AIDA64 results which are all in line with Phenom II @ ~3GHz. So this Llano sample has some "brake" in BIOS that is limiting your multiplier to 15 or 15.5x,even if you manually select 54x. So to recap everything: -Llano's turbo mode is broken on your sample, -real clock is 1.75x lower than what CPUz reports(2.4Ghz->1.4GHz,4.7Ghz->2.68Ghz,5.4Ghz->3.05Ghz) -results are in line with Phenom II ,clock per clock,give or take a few percent.

Just wanted to point out this obvious thing to all of you,so you all don't think Llano is slower than K7 at the same clock speed"
 
Just wanted to point out this obvious thing to all of you,so you all don't think Llano is slower than K7 at the same clock speed"

Gee..thanks for being obvious. It's as if we were all supposed to wade through some large thread that the OP didn't link to in order to find out some 'obvious' bit of information.

:thumbsdown:
 
Gee..thanks for being obvious. It's as if we were all supposed to wade through some large thread that the OP didn't link to in order to find out some 'obvious' bit of information.

:thumbsdown:

He should have put a quote block around it, but that was what the poster said on the other forum, he wasn't saying that to us :biggrin:
 
Argg i'm waiting for something that actually runs the way its supposed to? These early samples are in pretty bad shape.
 
33m9bva.jpg


2w3w5xi.jpg
 
Last edited:
FYI- The page that Grimpr's shots come from the 3.7 is the stable over air overclock the user got. There is also a naked unedited shot of the CPU with full model number.
 
Wow, that is super slow SuperPI. Just did couple calculation for the heck of it and got 40.xx for the first try and 37.xx for the second try. And also have a virtual machine running while that happens.

By the way, is the graphics part fast? Been a while since reading the 3DMark scores and having hard time interpreting them.
 
Wow, that is super slow SuperPI. Just did couple calculation for the heck of it and got 40.xx for the first try and 37.xx for the second try. And also have a virtual machine running while that happens.

By the way, is the graphics part fast? Been a while since reading the 3DMark scores and having hard time interpreting them.

I think it's basically at a Radeon HD 4830 level.
 
I think it's basically at a Radeon HD 4830 level.

Should hover around the Radeon 5550-5570 level for the top end Llano. Has 400 stream processors but they are clocked a bit lower at stock it seems not to mention having to use system memory.
 
Back
Top