AMD Myrtle engineering sample on Geekbench

[Slaughterem]

http://browser.primatelabs.com/geekbench3/search?utf8=%E2%9C%93&q=amd+myrtle

Not sure if this is Bristol Ridge or Summit Ridge (Zen) but comes up AMD Eng Sample: 2D3601AGM44AB_40/36/13/08_9874 @ 2.00 GHz
1 processor, 2 cores, 4 threads. Looks like Zen

Has 2 entries with scores of 2594 and 2679 for single core and 7985 and 8105 for multi core.

 

[Arachnotronic]

http://browser.primatelabs.com/geekbench3/search?utf8=%E2%9C%93&q=amd+myrtle

Not sure if this is Bristol Ridge or Summit Ridge (Zen) but comes up AMD Eng Sample: 2D3601AGM44AB_40/36/13/08_9874 @ 2.00 GHz
1 processor, 2 cores, 4 threads. Looks like Zen

Has 2 entries with scores of 2594 and 2679 for single core and 7985 and 8105 for multi core.

This is Bristol Ridge @ 4GHz.

 

[Slaughterem]

This is Bristol Ridge @ 4GHz.

How can you tell? I am not that familiar with Geekbench, Bristol Ridge on Geekbench is showing up as AMD FX-9800P RADEON R7, 12 COMPUTE CORES 4C+8G @ 1.85 GHz
1 processor, 4 cores I don't think that it would show as 2 core 4 thread but maybe I am wrong.

 

[Arachnotronic]

How can you tell? I am not that familiar with Geekbench, Bristol Ridge on Geekbench is showing up as AMD FX-9800P RADEON R7, 12 COMPUTE CORES 4C+8G @ 1.85 GHz
1 processor, 4 cores I don't think that it would show as 2 core 4 thread but maybe I am wrong.

L1 cache sizes are the giveaway. Also, that multi-core scaling is off the charts...>>>2x the ST score.

 

[MrTeal]

Is Zen still going to be Family 21?

 

[yuri69]

Is Zen still going to be Family 21?

No, Zen is called Family 17h. 17h in decimal equals 23. Thus you have to look for AMD Family 23.

Family 21 stands for 15h aka Bulldozer line.

 

[Slaughterem]

No, Zen is called Family 17h. 17h in decimal equals 23. Thus you have to look for AMD Family 23.

Family 21 stands for 15h aka Bulldozer line.

So this is excavator at 2 ghz 4 cores which is close to Intel Core i7-4750HQ
2000 MHz (4 cores) single core score 2642. Or how do you draw a comparison to an Intel mobile CPU?

 

[Arachnotronic]

So this is excavator at 2 ghz 4 cores which is close to Intel Core i7-4750HQ
2000 MHz (4 cores) single core score 2642. Or how do you draw a comparison to an Intel mobile CPU?

It's not actually at 2GHz, this is a bug in GB3's read of the frequency.

 

[Slaughterem]

It's not actually at 2GHz, this is a bug in GB3's read of the frequency.

Well that sucks I guess we have to wait until they launch at end of May and some laptops come out to really know what improvements BR brings.

 

[The Stilt]

what improvements BR brings.

None, as it is the same exact thing as Carrizo. Even the die shot released by AMD reveals that.

 

[Arachnotronic]

Well that sucks I guess we have to wait until they launch at end of May and some laptops come out to really know what improvements BR brings.

Bristol Ridge is just Carrizo for desktops. Go look at performance numbers for Carrizo and use your imagination to figure out how it would perform at higher clock-speeds

 

[Azuma Hazuki]

Works fine for low-mid-range builds. I don't give people i3's; A8-7600 (and now whatever the BR equivalent is) make a lot more sense. Now when they need more grunt, then I start tapping Intel for i5s and such. But below that? Intel's actually a pretty crap value.

 

[yuri69]

Bristol Ridge is just Carrizo for desktops.

It is Carrizo silicon with a new firmware. Various tweaks (mainly for better binning) could give up to 15% perf vs Carrizo, according to AMD's ISSCC paper: Increasing the performance of a 28nm x86-64 microprocessor through system power management

Basically, this silicon has now ability to achieve a bit higher frequencies at the same TDP than it would with the Carrizo's firmware.

 

[monstercameron]

It is Carrizo silicon with a new firmware. Various tweaks (mainly for better binning) could give up to 15% perf vs Carrizo, according to AMD's ISSCC paper: Increasing the performance of a 28nm x86-64 microprocessor through system power management

Basically, this silicon has now ability to achieve a bit higher frequencies at the same TDP than it would with the Carrizo's firmware.

Those kinda results can pribably only be seen in labs. Irl nigh impossible to tell.