Overclock.net › Forums › Industry News › Technology and Science News › [MacRumors] iPhone 5 Benchmarks Appear in Geekbench Showing a Dual Core, 1GHz A6 CPU
New Posts  All Forums:Forum Nav:

[MacRumors] iPhone 5 Benchmarks Appear in Geekbench Showing a Dual Core, 1GHz A6 CPU - Page 18

post #171 of 418
Quote:
Originally Posted by j3st3r View Post

Too bad it still gets smoked by an 'inferior phone'
Android noobs have no bounds to the limits of their delusional fantasies.

Please explain how the phone with the lower score (by over 200 points) is smoking the phone with the higher score. Your logic hurts my head. rolleyes.gif
Dominus Noctis
(33 items)
 
Sager NP7352
(14 items)
 
 
CPUMotherboardGraphicsRAM
[Intel] i7-4790K @ 4.5GHz [Gigabyte] Z97X-Gaming GT [EVGA] GeForce GTX 980 Ti Superclocked+ ACX 2.0... [G.SKILL] Ripjaws 16GB (4x4GB) @ 1600MHz 
Hard DriveHard DriveOptical DriveCooling
[Samsung] 840 Pro 256GB (OS/Programs/Games) [Western Digital] Caviar Blue 640GB (User Files) [HP] DVD-RW Multi-Recorder Lightscribe [Corsair] H100i Hydro Cooler 
CoolingCoolingCoolingCooling
[Corsair] SP120L High Performance 120mm Radiato... [Corsair] SP120L High Performance 120mm Radiato... [Corsair] SP120 High Performance Edition 120mm ... [Corsair] SP120 High Performance Edition 120mm ... 
CoolingCoolingCoolingOS
[Corsair] 140mm Top Exhaust Fan (White LED) [Corsair] 140mm Top Exhaust Fan (White LED) [Corsair] 140mm Rear Exhaust Fan [Microsoft] Windows 10 Pro (64-Bit) 
MonitorMonitorMonitorKeyboard
[ASUS] VG278HE 27' LED - 1920x1080p@144Hz [ASUS] VE278Q 27' LED - 1920x1080p@60Hz [Acer] H233H 23' LCD - 1920x1080p@30Hz [Corsair] Vengeance K70 (Silver) 
PowerCaseMouseMouse Pad
[Seasonic] X-1250W [Corsair] Carbide Series Air 540 (White) [RAZER] DeathAdder 3.5G [Corsair] MM200 Compact 
AudioAudioAudioAudio
[Schiit] Bifrost (Uber) Digital To Analog Conve... [Schiit] Lyr Headphone Amplifier [Focusrite] Scarlett 2i2 Audio Interface [Sennheiser] HD 650 Headphones 
AudioAudioAudioAudio
[V-MODA] Crossfade M-100 Headphones (Shadow) [KRK] Rokit 6 Studio Monitors [Samson] Q2U Microphone [Novation] Launchkey 25 MIDI Controller 
Other
[Linksys] AE3000 N900 USB Dual Band WiFi Adapter 
CPUMotherboardGraphicsGraphics
[Intel] Core i7-4800MQ @ 2.7GHz [Intel] HM87 Express Chipset [Nvidia] GeForce GTX 765M [Intel] HD Graphics 4600 
RAMHard DriveHard DriveOptical Drive
[Patriot] 16GB (2x8GB) @ 1600MHz [Intel] 525 Series 120GB (OS/Programs) [Seagate] Momentus 750GB (User Files/Games) [Toshiba/Samsung] 8X DVD±R/RW/4X +DL Super-Mult... 
CoolingOSMonitorMouse
[Cooler Master] NotePal U2 PLUS [Microsoft] Windows 10 Home (64-Bit) [Sager] 15.6' LED Display - 1920x1080p@60Hz [Logitech] Anywhere Mouse MX 
AudioAudio
[FiiO] E07K Andes USB DAC/AMP  [V-MODA] Crossfade M-100 Headphones (Shadow) 
  hide details  
Reply
Dominus Noctis
(33 items)
 
Sager NP7352
(14 items)
 
 
CPUMotherboardGraphicsRAM
[Intel] i7-4790K @ 4.5GHz [Gigabyte] Z97X-Gaming GT [EVGA] GeForce GTX 980 Ti Superclocked+ ACX 2.0... [G.SKILL] Ripjaws 16GB (4x4GB) @ 1600MHz 
Hard DriveHard DriveOptical DriveCooling
[Samsung] 840 Pro 256GB (OS/Programs/Games) [Western Digital] Caviar Blue 640GB (User Files) [HP] DVD-RW Multi-Recorder Lightscribe [Corsair] H100i Hydro Cooler 
CoolingCoolingCoolingCooling
[Corsair] SP120L High Performance 120mm Radiato... [Corsair] SP120L High Performance 120mm Radiato... [Corsair] SP120 High Performance Edition 120mm ... [Corsair] SP120 High Performance Edition 120mm ... 
CoolingCoolingCoolingOS
[Corsair] 140mm Top Exhaust Fan (White LED) [Corsair] 140mm Top Exhaust Fan (White LED) [Corsair] 140mm Rear Exhaust Fan [Microsoft] Windows 10 Pro (64-Bit) 
MonitorMonitorMonitorKeyboard
[ASUS] VG278HE 27' LED - 1920x1080p@144Hz [ASUS] VE278Q 27' LED - 1920x1080p@60Hz [Acer] H233H 23' LCD - 1920x1080p@30Hz [Corsair] Vengeance K70 (Silver) 
PowerCaseMouseMouse Pad
[Seasonic] X-1250W [Corsair] Carbide Series Air 540 (White) [RAZER] DeathAdder 3.5G [Corsair] MM200 Compact 
AudioAudioAudioAudio
[Schiit] Bifrost (Uber) Digital To Analog Conve... [Schiit] Lyr Headphone Amplifier [Focusrite] Scarlett 2i2 Audio Interface [Sennheiser] HD 650 Headphones 
AudioAudioAudioAudio
[V-MODA] Crossfade M-100 Headphones (Shadow) [KRK] Rokit 6 Studio Monitors [Samson] Q2U Microphone [Novation] Launchkey 25 MIDI Controller 
Other
[Linksys] AE3000 N900 USB Dual Band WiFi Adapter 
CPUMotherboardGraphicsGraphics
[Intel] Core i7-4800MQ @ 2.7GHz [Intel] HM87 Express Chipset [Nvidia] GeForce GTX 765M [Intel] HD Graphics 4600 
RAMHard DriveHard DriveOptical Drive
[Patriot] 16GB (2x8GB) @ 1600MHz [Intel] 525 Series 120GB (OS/Programs) [Seagate] Momentus 750GB (User Files/Games) [Toshiba/Samsung] 8X DVD±R/RW/4X +DL Super-Mult... 
CoolingOSMonitorMouse
[Cooler Master] NotePal U2 PLUS [Microsoft] Windows 10 Home (64-Bit) [Sager] 15.6' LED Display - 1920x1080p@60Hz [Logitech] Anywhere Mouse MX 
AudioAudio
[FiiO] E07K Andes USB DAC/AMP  [V-MODA] Crossfade M-100 Headphones (Shadow) 
  hide details  
Reply
post #172 of 418
Assuming that the Apple Geekbench score is accurate (and can't be compared to Android phone benchmarks). The results are still being inflated quite a bit by the memory bandwidth. Anand suggests that the A6 stuck with 1MB of cache because most iphone datasets didn't need more. If that is the case, then all the points gained by adding memory bandwidth don't really matter in real-world usage. That said, those are some healthy increases in theoretical performance.
post #173 of 418
Thread Starter 
Quote:
Originally Posted by bengal View Post

Well the benchmarks in the OP are legit. Engadget released their review of the iPhone 5. They got a score of 1628 in GeekBench. They got 924ms in Sunspider, which is faster than both the GSIII and HTC one X.
Quote:
Two times faster? Twice the graphics performance? Better battery life? Actually, yes. The iPhone 5 over-delivers on all those promises. Running the Geekbench test suite on the iPhone 4S gave us an average score of 634. The iPhone 5 netted an average of 1,628. That's more than twice as fast and, while you won't necessarily see such huge increases in day-to-day usage, apps do load noticeably quicker, HDR images are processed in half the time and tasks like video rendering in iMovie are equally expedient.

SunSpider scores average at 924ms, which is more than twice as fast as the 2,200ms the iPhone 4S manages and still quite a bit quicker than the 1,400ms scored by the Galaxy S III and the 1,700ms managed by the HTC One X. More important than numbers, web pages load very quickly, snapping into view as fast as your data plan can shovel the bits into Safari and, once there, smoothly reacting to your gestures.

Source
Ha, that's sweet. On iOS 6 the 4S would sit around 1800 on SunSpider so results of 900-1000 fit perfectly. Apparently frame rates are doubled too...
Kasuf
(9 items)
 
  
CPUMotherboardGraphicsRAM
Core i5 6600K ASRock Z170 Pro4 ASUS Radeon RX 480 ROG Strix Corsair Vengeance LPX 32GB 3000MHz 
Hard DriveCoolingMonitorPower
Samsung 850 EVO Noctua NH-D15 LG 34" Ultrawide (LG34UC98) Corsair HX750i 
Case
Silverstone FT05B-W 
  hide details  
Reply
Kasuf
(9 items)
 
  
CPUMotherboardGraphicsRAM
Core i5 6600K ASRock Z170 Pro4 ASUS Radeon RX 480 ROG Strix Corsair Vengeance LPX 32GB 3000MHz 
Hard DriveCoolingMonitorPower
Samsung 850 EVO Noctua NH-D15 LG 34" Ultrawide (LG34UC98) Corsair HX750i 
Case
Silverstone FT05B-W 
  hide details  
Reply
post #174 of 418
Quote:
Originally Posted by hajile View Post

Assuming that the Apple Geekbench score is accurate (and can't be compared to Android phone benchmarks). The results are still being inflated quite a bit by the memory bandwidth. Anand suggests that the A6 stuck with 1MB of cache because most iphone datasets didn't need more. If that is the case, then all the points gained by adding memory bandwidth don't really matter in real-world usage. That said, those are some healthy increases in theoretical performance.

(Emphasis mine)

Are you stating that based on your experience with mobile workloads?

ARM designs have historically been marred by a subpar memory controller. Improving that as much Apple did could potentially create a 1-to-1 percentage performance improvement in real world situations, as it would remove an old bottleneck.
post #175 of 418
I don't care for these random benchmarks. I am really happy with my S II and have no interest in getting a new phone for at least two years to come. I don't know what you guys do with your phones that you need quad cores running an OS that is mostly single threaded. Bar games I can't see these chips actually doing anything, they certainly can't fold or run Boinc.
The girlfriend.
(15 items)
 
The Mistress
(13 items)
 
Media Server
(11 items)
 
CPUMotherboardGraphicsRAM
FX8320 @ 4.5 Crosshair V (Third One) 290 Tri-X OC 16GB HyperX 
Hard DriveHard DriveHard DriveCooling
Sandisk SSD Plus 240GB Toshiba 3TB Hitachi 2TB Swiftech H220-X 
OSMonitorKeyboardPower
Win 8.1/Mint Viewsonic VX2457-mhd Fuhlen L411 Slim Super Flower Golden King 650W 
CaseMouseMouse Pad
Enthoo Pro G400 Super Flower 
CPUMotherboardGraphicsRAM
A8-6410 Lenovo Lancer 4B2 K16.3 R5 128 Shaders/M230 Hynix 8GB DDR3 1600 
Hard DriveHard DriveOSMonitor
Samsung 840 120 GB SSD Seagate Momentus 1TB 5400rmp Win 8.1 CMN1487 TN LED 14" 1366*768 
KeyboardPowerMouseMouse Pad
Lenovo AccuType 2900mAh/41Wh Elan Trackpad/Logitech M90 Super Flower 
Audio
AMD Avalon(Connexant) 
  hide details  
Reply
The girlfriend.
(15 items)
 
The Mistress
(13 items)
 
Media Server
(11 items)
 
CPUMotherboardGraphicsRAM
FX8320 @ 4.5 Crosshair V (Third One) 290 Tri-X OC 16GB HyperX 
Hard DriveHard DriveHard DriveCooling
Sandisk SSD Plus 240GB Toshiba 3TB Hitachi 2TB Swiftech H220-X 
OSMonitorKeyboardPower
Win 8.1/Mint Viewsonic VX2457-mhd Fuhlen L411 Slim Super Flower Golden King 650W 
CaseMouseMouse Pad
Enthoo Pro G400 Super Flower 
CPUMotherboardGraphicsRAM
A8-6410 Lenovo Lancer 4B2 K16.3 R5 128 Shaders/M230 Hynix 8GB DDR3 1600 
Hard DriveHard DriveOSMonitor
Samsung 840 120 GB SSD Seagate Momentus 1TB 5400rmp Win 8.1 CMN1487 TN LED 14" 1366*768 
KeyboardPowerMouseMouse Pad
Lenovo AccuType 2900mAh/41Wh Elan Trackpad/Logitech M90 Super Flower 
Audio
AMD Avalon(Connexant) 
  hide details  
Reply
post #176 of 418
Quote:
Originally Posted by lvlrdka22 View Post

Methinks this article is a ploy to get more sales on the benchmark app.
1. Post (real or fake) benchmark showing Android devices being slower than iDevices.
2. Android fan(boys) rage.
3. Android fans hit the store and buy the app to attempt to disprove the benchmark.
4. ???
5. Profit!!!
Note: doesn't work on Apple fan(boys) because they just don't care unless they're winning the benchmark (no offense to them tongue.gif).

this at .99 a pop exactly
Easy V3
(17 items)
 
12 Pack
(15 items)
 
 
CPUMotherboardGraphicsRAM
i7-4820k Asus Rampage IV BE evga 1080ti FE w/ hybrid kit 16 GB G.Skill Trident X 2400mhz  
Hard DriveHard DriveOptical DriveCooling
Samsung 840 EVO 120gb Samsung 840 EVO 1TB Samsung Blu-ray Corsair H110 
OSMonitorKeyboardPower
windows 10 Asus PG278Q Ducky Shine 3 - blue led - cherry mx browns Seasonic Platinum-1000 
CaseMouseMouse PadAudio
Dimastech Easy V3 Mionix NAOS 8200 Mionix Ensis 320 titanium HD  
Other
audio technica ATH-A900X 
  hide details  
Reply
Easy V3
(17 items)
 
12 Pack
(15 items)
 
 
CPUMotherboardGraphicsRAM
i7-4820k Asus Rampage IV BE evga 1080ti FE w/ hybrid kit 16 GB G.Skill Trident X 2400mhz  
Hard DriveHard DriveOptical DriveCooling
Samsung 840 EVO 120gb Samsung 840 EVO 1TB Samsung Blu-ray Corsair H110 
OSMonitorKeyboardPower
windows 10 Asus PG278Q Ducky Shine 3 - blue led - cherry mx browns Seasonic Platinum-1000 
CaseMouseMouse PadAudio
Dimastech Easy V3 Mionix NAOS 8200 Mionix Ensis 320 titanium HD  
Other
audio technica ATH-A900X 
  hide details  
Reply
post #177 of 418
Quote:
Originally Posted by mavere View Post

(Emphasis mine)
Are you stating that based on your experience with mobile workloads?
ARM designs have historically been marred by a subpar memory controller. Improving that as much Apple did could potentially create a 1-to-1 percentage performance improvement in real world situations, as it would remove an old bottleneck.

Most computer calculations for any particular problem/subroutine occur in close memory proximity. For this reason, processors take everything near where they're working and cache it in the processor's L1/L2/L3 (with each representing stuff farther away and thus less likely to be requested). When the processor needs a piece of data, it searches each cache (the caches being arranged as associated arrays). If the computer put the right data in the cache, it can pretend that it's accessing main memory (actually, in some cases such as x86, the architecture is so old that the ISA doesn't account for cache so designers have to put the cache in so that the processor doesn't realize that the cache exists), but saves huge amounts of time not having to wait for main memory to respond (and memory speed and bandwidth don't matter in this case since they aren't being used). Since most reads and writes occur in close proximity, this works well.

One problem that occurs is having too large a dataset to fit in cache. If a chip has 1MB of cache, but the working set (ie, the group of data and instructions the computer is reading/writing) is larger than 1MB, then lots of time is spent waiting on RAM (and also on changing out the cache ("cache thrashing") as the processor attempts to cache what is necessary despite the fact that it doesn't physically have the room).

Now, let's assume that Anand is right and Apple didn't change the cache size because most iOS workloads fit inside 1MB. If this is true, then memory bandwidth isn't the biggest concern as the data can be gradually changed in and out while most of the read/write activity is being done inside the processor. This principle is the same whether the processor is a microcontroller or a supercomputer. Being a mobile workload doesn't change this.

The only question is whether or not Apple chose to leave the cache alone for this reason. Anand believes this is the reason (only Apple would truly know and they probably won't ever say), so I took what he said in the spirit he intended (that is, an educated guess (not definitive) based on years doing what he does and his having an EE degree to back some of that up even more).
Edited by hajile - 9/18/12 at 9:27pm
post #178 of 418
Just thought I leave it here

49915.png

http://www.anandtech.com/show/6309/iphone-5a6-sunspider-performance
Quote:
Originally Posted by Tman5293 View Post

Please explain how the phone with the lower score (by over 200 points) is smoking the phone with the higher score. Your logic hurts my head. rolleyes.gif

Look above.
Edited by bengal - 9/18/12 at 11:04pm
post #179 of 418
Susncript is better than the GNote II... drool.gif
post #180 of 418
Quote:
Originally Posted by bengal View Post

Well the benchmarks in the OP are legit. Engadget released their review of the iPhone 5. They got a score of 1628 in GeekBench. They got 924ms in Sunspider, which is faster than both the GSIII and HTC one X.
Source
Well I'll be. I didn't think they'd do it, but they did. The sunspider score is also lower than the Atom powered phone!
Quote:
Originally Posted by hajile View Post

Assuming that the Apple Geekbench score is accurate (and can't be compared to Android phone benchmarks). The results are still being inflated quite a bit by the memory bandwidth. Anand suggests that the A6 stuck with 1MB of cache because most iphone datasets didn't need more. If that is the case, then all the points gained by adding memory bandwidth don't really matter in real-world usage. That said, those are some healthy increases in theoretical performance.
Almost everything will benefit from the much improved memory interface. It's a big reason the Atom outperforms A9 SoCs.
:O
(12 items)
 
  
CPUMotherboardGraphicsRAM
i7-3930k Rampage IV Formula ASUS HD 7970 DCII 16GB G.Skill 1600Mhz 
Hard DriveHard DriveCoolingOS
1.5TB WD Caviar Black 64GB Crucial M4 Corsair H100i Windows 7 
MonitorPowerCaseAudio
Asus 24" 1920x1080 Thermaltake 750w NZXT Phantom Orange (special edition!) Sennheiser HD 280 
  hide details  
Reply
:O
(12 items)
 
  
CPUMotherboardGraphicsRAM
i7-3930k Rampage IV Formula ASUS HD 7970 DCII 16GB G.Skill 1600Mhz 
Hard DriveHard DriveCoolingOS
1.5TB WD Caviar Black 64GB Crucial M4 Corsair H100i Windows 7 
MonitorPowerCaseAudio
Asus 24" 1920x1080 Thermaltake 750w NZXT Phantom Orange (special edition!) Sennheiser HD 280 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Technology and Science News
Overclock.net › Forums › Industry News › Technology and Science News › [MacRumors] iPhone 5 Benchmarks Appear in Geekbench Showing a Dual Core, 1GHz A6 CPU