Kabini IPC, what generation are we talking about?

[monstercameron]

So it would be pointless to upgrade my friend with an Athlon64 X2 4800+ S939 rig with 4x512MB DDR, to a 5350 AM1 rig then? Seeing as how his IPC and clockspeed would be going down?

more cores and lower powerdraw means it would be worthwhile.

 

[Tsavo]

So it would be pointless to upgrade my friend with an Athlon64 X2 4800+ S939 rig with 4x512MB DDR, to a 5350 AM1 rig then? Seeing as how his IPC and clockspeed would be going down?

Utterly pointless next to all the cheap OCable Pentium deals out there.

 

[VirtualLarry]

Utterly pointless next to all the cheap OCable Pentium deals out there.

That's a big IF, if I can manage to catch a combo deal for around $100 or less, for overclockable mobo and G3258. The AM1 mobo + quad-core CPU are around that price normally, no special deals needed.

 

[Tsavo]

That's a big IF, if I can manage to catch a combo deal for around $100 or less, for overclockable mobo and G3258. The AM1 mobo + quad-core CPU are around that price normally, no special deals needed.

http://forums.anandtech.com/showthread.php?t=2393267

:whiste:

:biggrin:

 

[Roland00Address]

So it would be pointless to upgrade my friend with an Athlon64 X2 4800+ S939 rig with 4x512MB DDR, to a 5350 AM1 rig then? Seeing as how his IPC and clockspeed would be going down?

For roughly the same price you can get a pentium g3258 or a similar pentium or celeron. Here is the 5350 compared to the g3420 which is the same ghz, cache and generation as the g3258

note I did the percents how much one processor is faster than another. Thus a 0% means they are the same speed, and a 100% means twice as fast, 200% is three times as fast etc.

http://www.anandtech.com/bench/product/1256?vs=1223
Intel G3420 vs AMD 5350

Intel Wins
% won by Highter is Better and the Intel Pentium Wins
175.56% --- Cinebench R15 - Single Threaded --- Score in CBMarks - Higher is Better
_______ - - - 124 vs 45
_46.34% --- Cinebench R15 - Multi-Threaded --- Score in CBMarks - Higher is Better
_______ - - - 240 vs 164
_99.46% --- 3D Particle Movement: Single Threaded --- Score (Higher is Better) 106.23 vs 53.26
_______ - - -
166.79% --- WebXPRT --- Score (Higher is Better)
_______ - - - 2241 vs 840
210.06% --- Google Octane v2 --- Score (Higher is Better)
_______ - - - 24349 vs 7853
178.90% --- SYSmark 2014 - Office Productivity --- Score in SYSmarks (Higher is Better)
_______ - - - 1163 vs 417
166.17% --- SYSmark 2014 - Media Creation --- Score in SYSmarks (Higher is Better) 1062 vs 399
_______ - - -
105.81% --- SYSmark 2014 - Data and Financial Analysis --- Score in SYSmarks (Higher is Better)
_______ - - - 1099 vs 534
148.21% --- SYSmark 2014 - Overall --- Score in SYSmarks (Higher is Better)
_______ - - - 1107 vs 446
205.64% --- Cinebench R10 - Single Threaded Benchmark --- Score in CBMarks - Higher is Better
_______ - - - 6235 vs 2040
_70.14% --- Cinebench R10 - Multi-Threaded Benchmark --- Score in CBMarks - Higher is Better
_______ - - - 11869 vs 6976
157.41% --- Cinebench 11.5 - Single Threaded --- Score in CBMarks - Higher is Better
_______ - - - 1.39 vs 0.54
_34.31% --- Cinebench 11.5 - Multi-Threaded --- Score in CBMarks - Higher is Better
_______ - - - 2.74 vs 2.04
_82.01% --- x264 HD Benchmark - 1st pass - v3.03 --- Frames per Second - Higher is Better
_______ - - - 65.87 vs 36.19
_52.47% --- x264 HD Benchmark - 2nd pass - v3.03 --- Frames per Second - Higher is Better
_______ - - - 16.97 vs 11.13
_31.49% --- 7-zip Benchmark --- 32MB Dictionary - Total MIPS - Higher is Better
_______ - - - 7746 vs 5891

Intel Wins
% won by Lower is Better and the Intel Pentium Wins
325.95% --- Dolphin Emulation Benchmark --- Time in Minutes (Lower is Better)
_______ - - - 9.48 vs 40.38
130.55% --- WinRAR 5.01, 2867 files, 1.52 GB --- Time in Seconds (Lower is Better)
_______ - - - 141.36 vs 325.91
133.96% --- FastStone Image Viewer 4.9 --- Time in Seconds (Lower is Better)
_______ - - - 53 vs 124
180.26% --- Sunspider 1.0.2 --- Time in Milliseconds (Lower is Better)
_______ - - - 152 vs 426
244.55% --- Kraken 1.1 --- Time in Milliseconds (Lower is Better)
_______ - - - 1322 vs 4555


Amd Wins
% won by Higher is better and the AMD Athlon Quad Core Wins
__1.33% --- 3D Particle Movement: MultiThreaded --- Score (Higher is Better)
_______ - - - 171.79 vs 174.07
301.34% --- AES-128 Performance - TrueCrypt 7.1 Benchmark --- Mean Encryption/Decryption AES Algorithm - GB/s
_______ - - - 0.299 vs 1.2

As you can see unless you are doing encryption and decryption the intel is just a much better processor.

 

[VirtualLarry]

http://forums.anandtech.com/showthread.php?t=2393267

:whiste:

:biggrin:

LOL. Yes, I know I posted that, but I won't have money till Fri, and that mobo has some really poor reviews.

 

[Exophase]

Bobcat is around P4 IPC:
http://anandtech.com/bench/product/328?vs=92

.. how could you possibly draw that conclusion from the link you gave? This is a list of perf for Bobcat vs P4 normalized vs MHz:

1.1766146993 1.973791894 1.4134615385 1.4153225806 1.2469879518 1.7835365854 1.5 1.7360022396 2.0208643815 2.2660714286 2.25 1.772260274 1.6814079422 2.25 1.8820443751 1.9571154041 2.1934501142 2.2535944744 2.2131610707 2.3972873425 2.3691401329 1.4782721159 1.7729079894 2.7946963216 1.821557971 1.5719594595

It only comes close to 1.0 for the very first test, and even then not very close. Average is 1.89. Definitely not around the same at all.

 

[ShintaiDK]

.. how could you possibly draw that conclusion from the link you gave? This is a list of perf for Bobcat vs P4 normalized vs MHz:



It only comes close to 1.0 for the very first test, and even then not very close. Average is 1.89. Definitely not around the same at all.

Did you exclude all the multithreaded benches, since the bobcat is a dualcore and the P4 a singlecore?

 

[ShintaiDK]

So it would be pointless to upgrade my friend with an Athlon64 X2 4800+ S939 rig with 4x512MB DDR, to a 5350 AM1 rig then? Seeing as how his IPC and clockspeed would be going down?

Utter waste of money compared to a Celeron/Pentium.

AM1 platform offers absolutely nothing.

 

[VirtualLarry]

AM1 platform offers absolutely nothing.

Lower power? Lower cost? You sure about that?

 

[ShintaiDK]

Lower power? Lower cost? You sure about that?

A fraction of the performance. The saved cost in the mobo you add to the CPU. With a cheap LGA1150 board its actually cheaper with CPU than the AM1.(40+50 vs 33+65 using newegg.)

Lower power? Not really.

Its a lose/lose.

 

[monstercameron]

what skus are you comparing?

 

[AtenRa]

A fraction of the performance. The saved cost in the mobo you add to the CPU. With a cheap LGA1150 board its actually cheaper with CPU than the AM1.(40+50 vs 33+65 using newegg.)

Lower power? Not really.

So according to Newegg prices, the TOP-end Kabini + mobo is cheaper than the entry level Celeron.

Lower power? Not really.

45% more power usage in CPU + GPU load for the Celeron G1820 than TOP-end Athlon 5350

http://www.techspot.com/review/806-amd-kabini-vs-intel-bay-trail-d/page8.html

 

[ShintaiDK]

So according to Newegg prices, the TOP-end Kabini + mobo is cheaper than the entry level Celeron.

No it cost more.

45% more power usage in CPU + GPU load for the Celeron G1820 than TOP-end Athlon 5350

http://www.techspot.com/review/806-amd-kabini-vs-intel-bay-trail-d/page8.html

The Celeron will complete the task much much faster and return to idle. it will also be able to substain idle speeds at light loads unlike the Kabini. Alternatively you could just get the T model Celeron as well.

Not to mention the Celeron idles at lower power consumption.

Kabini is a lose/lose product on AM1.

 

[Ancalagon44]

I haven't read the whole thread yet, I will later, but it's generally agreed that Kabini offers 90% of IPC of K8.(which is coincidentally the official number from AMD) K8, on the other hand has 25% lower IPC then C2D, which has 96% higher IPC then P4. Kaveri/Puma offered 20% more IPC then Kabini so it clearly surpassed K8 and came close to the original Phenom. Saying that Kabini has the same IPC as P4 is just wrong.

Kaveri != Puma

 

[NTMBK]

A fraction of the performance. The saved cost in the mobo you add to the CPU. With a cheap LGA1150 board its actually cheaper with CPU than the AM1.(40+50 vs 33+65 using newegg.)

Lower power? Not really.

Its a lose/lose.

On eBuyer (UK site) the cheapest Haswell Celeron is £29.99 and the cheapest LGA1150 motherboard is £31.23. The Athlon 5350 is £39.99 and the cheapest AM1 motherboard is £21.87. The price difference is a whopping 64 pence. Wow. You could buy an entire can of Coke with that price difference! In return you get a significantly lower TDP (25W SoC vs. 53W APU + 4W chipset), a faster GPU, support for modern ISA extensions and better multithreaded performance- while giving up some single threaded performance.

 

[NTMBK]

Not to mention the Celeron idles at lower power consumption.

I don't know how Techspot got those numbers, because Bit-Tech list Athlon 5350 idle power as 16W http://www.bit-tech.net/hardware/2014/04/24/amd-athlon-5350-kabini-review/8

 

[ShintaiDK]

On eBuyer (UK site) the cheapest Celeron is £29.99 and the cheapest LGA1150 motherboard is £31.23. The Athlon 5350 is £39.99 and the cheapest AM1 motherboard is £21.87. The price difference is a whopping 64 pence. Wow. You could buy an entire can of Coke with that price difference! In return you get a significantly lower TDP (25W SoC vs. 53W APU + 4W chipset), a faster GPU, support for modern ISA extensions and better multithreaded performance- while giving up some single threaded performance.

Is the graphics faster? I would like to see that. Even AMD only dares to compare it to Baytrail graphics.

You can also replace the 53W with a 35W T model. not to mention again, the Celeron idles at lower power.

And in terms of multithreaded performance.
http://anandtech.com/bench/product/1223?vs=1257

And this is using a slower outdated Ivy Bridge Celeron.

Kabini is simply a lose/lose CPU. I guess you could somehow excuse it if you needed less than 35W TDP, but you could be up to 25W TDP.

 

[NTMBK]

Oh, and guru3D measured a 22W idle power. http://www.guru3d.com/articles_pages/amd_athlon_5350_apu_and_am1_platform_review,6.html

 

[ShintaiDK]

I don't know how Techspot got those numbers, because Bit-Tech list Athlon 5350 idle power as 16W http://www.bit-tech.net/hardware/2014/04/24/amd-athlon-5350-kabini-review/8

The site you link reused different numbers across different reviews. Their test setup only specifies the Kabini.

If you compare these 2 for example:
http://www.bit-tech.net/hardware/2014/04/24/amd-athlon-5350-kabini-review/8
http://www.bit-tech.net/hardware/cpus/2014/01/30/amd-a10-7850k-and-a10-7700k-kaveri-review/10

Same numbers, one uses a standard 750W PSU the other a super efficient 400W 80Plus gold.

 

[ShintaiDK]

Oh, and guru3D measured a 22W idle power. http://www.guru3d.com/articles_pages/amd_athlon_5350_apu_and_am1_platform_review,6.html

Also using:
400 Watt 80 Plus Gold.

I assume it was AMD supplied since there seem to be a trend.

And only listing the Kabini hardware as well.

 

[NTMBK]

Multithreaded performance.

 

[NTMBK]

Also using:
400 Watt 80 Plus Gold.

I assume it was AMD supplied since there seem to be a trend.

And only listing the Kabini hardware as well.

You use a low watt power supply for a low power APU. Heck, even 400W is probably too much; you'd be better off with a laptop power brick and one of the motherboards with a DC jack.

 

[ShintaiDK]

You use a low watt power supply for a low power APU. Heck, even 400W is probably too much; you'd be better off with a laptop power brick and one of the motherboards with a DC jack.

And you would too as well for any IGP platform if you wish to slow the lowest numbers. However since they dont, you cant compare the 2.

 

[ShintaiDK]

Multithreaded performance.

Explain to me why the i3 4330 gets roughly 1/3rd of the points in 3D particle vs the i3 3225. Else you perfectly show that the Haswell Celeron will be faster and prove my point. Thanks.