Intel Z3770 geekbench sighted

[shady28]

but thats just cores right? no north bridge, no GPU. jaguars cores for example are what 3.1mm sq.

Exactly, which is why I said your other points were dead spot on. The Nano needs all kinds of supporting chips. I think the Nano was more for a low footprint mini PC aka 'net top'. It is simply not a mobile chip and comparing it to these new Atoms isn't very meaningful, IMO. OTOH, for anyone getting a net top, this might be useful info (and they are still popular for things like kiosks and such).

 

[Nothingness]

Forget Geekbench, especially Floating point results. It is flawed.

Apparently only two of the FP benchmarks have an issue. For the rest noone has found issues yet and even though optimization was not pushed, the same level was applied to all platforms.

I think this is still the best CPU benchmark for mobile platforms even ifit's not perfect (and no benchmark ever will be perfect).

 

[Dresdenboy]

Apparently only two of the FP benchmarks have an issue. For the rest noone has found issues yet and even though optimization was not pushed, the same level was applied to all platforms.

I think this is still the best CPU benchmark for mobile platforms even ifit's not perfect (and no benchmark ever will be perfect).

We even might recalculate Geekbench scores leaving out the issue laden sub benches.

 

[galego]

Forget Geekbench, especially Floating point results. It is flawed.

No

 

[mikk]

No

Yes. Unreliable as hell. A bench with this flaws cannot be considered as a reliable benchmark anymore. Geekbench lost all its reputation. If you trust Geekbench you have to trust Antutu as well.

 

[galego]

Yes. Unreliable as hell. A bench with this flaws cannot be considered as a reliable benchmark anymore. Geekbench lost all its reputation. If you trust Geekbench you have to trust Antutu as well.

Geekbench: fair benchmark that needs improvement in some weak point.

Older Antutu: unfair benchmark used to fake Intel chip scores.
Modern Antutu: Older minus the fake stuff.

Geekbench reputation continues being in good shape.

 

[Nothingness]

Yes. Unreliable as hell. A bench with this flaws cannot be considered as a reliable benchmark anymore. Geekbench lost all its reputation. If you trust Geekbench you have to trust Antutu as well.

One has to like people who have no clue, it makes other seem brighter.

 

[jfpoole]

We even might recalculate Geekbench scores leaving out the issue laden sub benches.

I've uploaded a simple script (geekbench-recalculate.py) to our Geekbench tools repo that makes this easy: https://github.com/primatelabs/geekbench-tools/

Just point the script at the URL of the result you want to recalculate and it'll do the rest. Right now the script just excludes the Sharpen Image and the Blur Image workloads, but I've written it in such a way that it would be easy to modify to exclude other workloads as well.

Let me know if you have any questions or issues with the script.

 

[Dresdenboy]

I've uploaded a simple script (geekbench-recalculate.py) to our Geekbench tools repo that makes this easy: https://github.com/primatelabs/geekbench-tools/

Just point the script at the URL of the result you want to recalculate and it'll do the rest. Right now the script just excludes the Sharpen Image and the Blur Image workloads, but I've written it in such a way that it would be easy to modify to exclude other workloads as well.

Let me know if you have any questions or issues with the script.

Thank you. I'll give it a try.

 

[Dresdenboy]

Update: The script works for me. I think, it will help creating some analysis of the results and different platform conditions (e.g. 32 vs. 64 bit)

 

[mikk]

One has to like people who have no clue, it makes other seem brighter.

Of course, you are the perfect example.

 

[mikk]

Way too low scores for the Z3770 imho. This CPU should easily crush a 2 core 1,0 Ghz Temash.

I was right.

http://browser.primatelabs.com/geekbench2/2226937


These are better, but still unsure. With ES samples we never can be sure that it runs with nominal frequency. Furthermore Geekbench is unreliable and broken.

 

[TuxDave]

I was right.

http://browser.primatelabs.com/geekbench2/2226937


These are better, but still unsure. With ES samples we never can be sure that it runs with nominal frequency. Furthermore Geekbench is unreliable and broken.

And plus you have to ignore the blur and sharpen image fp scores until Geekbench 3 when it gets fixed.

 

[monstercameron]

looks good http://browser.primatelabs.com/geekbench2/compare/2119445/2226937
but I doubt that the 2W spd is anywhere accurate. but this does bode well for future atom tablets!

also vs the faster a4-5000 http://browser.primatelabs.com/geekbench2/compare/2219941/2226937

aside: maybe kde and gnome would love machines like these with foss intel gpu drivers in tablets!

 

[Nothingness]

I was right.

http://browser.primatelabs.com/geekbench2/2226937


These are better, but still unsure. With ES samples we never can be sure that it runs with nominal frequency.

Yup, you were right. If you scale the previous integer results, it looks like the 1.46 GHz was right and this new result would be 2.4GHz (one can assume linear scaling for single-threaded integer tasks in Geekbench). That seems to make sense, but this is not proven yet.

Hopefully Geekbench v3 will be ready for BT launch in September and will fix existing issues (the two buggy FP tests, and lack of compiler optimization for both Intel and ARM).

 

[d3m]

The Antutu benchmarks comparing Baytrail clocked at 1.4GHz (45,000) to Snapdragon 800 (30,000) i.e. the best SoC from ARM camp this year, shows an impressive lead of Baytrail over ARM, even when using an UNDERCLOCKED Baytrail

Later this year, we'll see the brainless tech blogging sheeps telling the exact opposite story from last year's "Intel is dead, ARM is king" nonsense to "Intel makes a comeback, challenge the ARM's throne". Go Intel!!!

Bwaahahhahhahahahahah AnTuTu :biggrin:

 

[sm625]

Blur Image
single-core scalar

Intel Celeron 1007U @ 1.50 GHz:    1379
AMD A6-1450 APU:                   3289
Intel Atom Z3770 @ 1.46 GHz:       870

And I'm supposed to believe that?

 

[mikk]

Some results are strange in Geekbench. In the past Geekbench favoured AMD over Intel. The reality was different. So it's a good sign when Bay Trail shows a good performance even in Geekbench.

 

[TuxDave]

Blur Image
single-core scalar

Intel Celeron 1007U @ 1.50 GHz:    1379
AMD A6-1450 APU:                   3289
Intel Atom Z3770 @ 1.46 GHz:       870

And I'm supposed to believe that?

I wouldn't believe anything under that specific test until Geekbench3.

 

[LegSWAT]

Oh, a secret 50% extra performance switch. Who made that up?

 

[Arachnotronic]

Oh, a secret 50% extra performance switch. Who made that up?

It's called Turbo.

 

[LegSWAT]

Like in the old days

Turbo

 

[MisterMac]

Oh, a secret 50% extra performance switch. Who made that up?

I'm sorry that' just made my day!
Hahaha - fantastic!

I imagine Otellini or the likes coming down to the boys in the lab headed up by Francois Piednoel.

"These results are horrible - the Investors and PR will kill us!"

FP: (In a super sexy french accent)
"Don't worry - we've not turned on the secret performance switch!".

*click*

 

[LegSWAT]

FP: (In a super sexy french accent)
"Don't worry - we've not turned on the secret performance switch!".

*click*

You're talking about a super sexy french accent that sounds just as super sexy as the maître d'hotel (played by John Cleese) serving Mr. Creusote in Monty Python's Meaning of Life??? :biggrin:

 

[Dresdenboy]

• Intel Atom Z3770 1.46GHz vs. Intel Atom Z3770 1.46GHz
GenuineIntel Family 6 Model 55 Stepping 2 vs. GenuineIntel Family 6 Model 55 Stepping 3

• Geekbench 2 Score: 2093 vs. 3093 •

• Integer Performance: 2353 vs. 3582
• Floating Point Performance: 2044 vs. 3140
• Memory Performance: 1761 vs. 2420
• Stream Performance: 2021 vs. 2572

source: http://browser.primatelabs.com/geekbench2/compare/2226937/2068301

If older scores are at 1.46GHz, the new sub scores were measured at 2.4GHz, some scaling losses included.