>AnandTech says that it's "quite astonishing" how close the A12 and the previous-generation A11 are to desktop CPUs, with "very small margins until Apple’s mobile SoCs outperform the fastest desktop CPUs in terms of ST performance."
>As part of the review, AnandTech also offered a look at how Apple has improved performance in older devices by tweaking scaling performance. The A9 in the iPhone 6s, for example, took 435ms for the CPU to reach maximum frequency, but that time was cut to 80ms in iOS 12 for a "great boost to performance in shorter interactive workloads."
>Apple's marketing department was really underselling the improvements here by just quoting 15% - a lot of workloads will be seeing performance improvements I estimate to be around 40%, with even greater improvements in some corner-cases. Apple's CPU have gotten so performant now, that we're just margins off the best desktop CPUs; it will be interesting to see how the coming years evolve, and what this means for Apple's non-mobile products.
>spend time boosting the performance of older phones >blow the fuck out of everyone else with each new CPU release >edging close to laptop performance in a phone
Is anyone else even trying? Androjeets/Samshills, what are your pathetic excuses going to be this time?
Did they fix mail notifications yet? Because for anyone that is getting 30+ mails daily, its fucking garbage. But oh HarDcorE RaCinG 365 is now 97fps instead of 83fps.
Hunter Perez
does that mean that i can run rpcs3, oh no wait yiu can only use useless trash from the fagstore.
Hunter Adams
No you can't run Raspberry Pi Counter Strike: 3 you dumb virgin.
Oliver Robinson
>Android: Toolchain: NDK r16 LLVM compiler, Flags: -Ofast, -mcpu=cortex-A53 >-mcpu=cortex-A53 The closest optimal mcpu target for kryo/sd835,845 is cortex-a57, cortex-a53 is for power efficiency and not performance, and yes it does make a huge difference. Not to mention that llvm is shit when it comes to performance. Rerun the test with linaro's gcc and target cortex-a57 instead.
In Xcode, Apple ships a closed-source version of clang, which in turn has a closed-source version of LLVM (it seems like Xcode 9's LLVM forks off of LLVM 5?). Apple's clang/LLVM generally includes stability enhancements that would otherwise potentially cause problems for developers. I'm not sure what the benefits are of using Rust's current LLVM, e.g. it can be newer than Apple's, but Apple's clang is a really safe bet and seems a good thing to use by default. Some of these platform-specific issues can manifest in weird and very difficult to diagnose ways, e.g. random restarts, that it would be good to save developers from.
github.com/llvm-mirror/llvm/blob/master/lib/Target/AArch64/AArch64SchedExynosM3.td def ExynosM3Model : SchedMachineModel { let IssueWidth = 6; // Up to 6 uops per cycle. let MicroOpBufferSize = 228; // ROB size. let LoopMicroOpBufferSize = 40; // Based on the instruction queue size. let LoadLatency = 4; // Optimistic load cases. let MispredictPenalty = 16; // Minimum branch misprediction penalty. let CompleteModel = 1; // Use the default model otherwise.
list UnsupportedFeatures = [HasSVE];
// FIXME: Remove when all errors have been fixed. let FullInstRWOverlapCheck = 0; } [\code] github.com/llvm-mirror/llvm/blob/master/lib/Target/AArch64/AArch64SchedA53.td def CortexA53Model : SchedMachineModel { let MicroOpBufferSize = 0; // Explicitly set to zero since A53 is in-order. let IssueWidth = 2; // 2 micro-ops are dispatched per cycle. let LoadLatency = 3; // Optimistic load latency assuming bypass. // This is overriden by OperandCycles if the // Itineraries are queried instead. let MispredictPenalty = 9; // Based on "Cortex-A53 Software Optimisation // Specification - Instruction Timings" // v 1.0 Spreadsheet let CompleteModel = 1;
>Recently designed Specialty CPUs with Reduced and Targeted Instruction Sets and MicroCode Optimizations >Similar performance to multipurpose, over-encompassing CPUs built to maintain compatibility with a 40-year old processor and instruction set, which has only gotten more bloated from 40-years of adding on to it
Honestly, the fact that Apple's Purpose Built Race Car of a CPU is only just now able to keep up with the giant, do anything, go anywhere lorry that is x86 CPUs in Single-Threaded instructions is kind of disappointing.
When GPUs changed hardware for DX10, and went to the massively-parallel approach, they were able to just man-handle CPUs easily in SIMD/SIMT calculations.
But nobody called for GPUs to replace CPUs because surprise-surprise, CPUs are still way faster at a lot of things, given their broad instruction-set. Apple's little ARM cores can benchmark well, because they're built to run that race, well. Hell, they may even approach traditional CPUs in raw FLOPs. But x86 CPUs will always have the computational advantage of having a colossal instruction-set to work with, making them ideal for more tasks, more often.
Mason Miller
> but still we’re now talking about very small margins until Apple’s mobile SoCs outperform the fastest desktop CPUs in terms of ST performance. Uhh where are the results or graphs for this? I don't see any benchmarks with a 'fastest desktop CPU' in that AnandTech article.
Jose Hernandez
AnandTech link to Anandtech test for server using Xeon platinum 8180 but Anandtech reviewer use very bad gcc compilator configuration over clasic ICC Intel and compare A12 single program per test over Xeon simultaneous program per thread.
You already got BTFO the last 50 times you shilled this Ranjeet. You masochistic or something? Applel is fucking trash and the only thing they know how to do is cheat at artificial benchmarks with hardware tailored to it. Too bad for you they can't cheat real world app performance.
Anandtech is one of, if not Jow Forums's most trusted hardware reviewer for actual technical information. You don't know this because you're a rebbitor. Why are you using PooOS?
Colton Thompson
>t. SEETHING iNigger
Mason Nelson
>iPhone 11 better than iPhone 6 >Haha! We win! Congratulations to all iPhone 11 owners for winning when vs iPhone 6 owners.
Owen Fisher
No one cares that your street shitting cousin works at Applel as a hired shill Dikshit.
>hates Indians >uses Windows and an Android phone >made Indian CEOs
I wonder what your agenda is. You're one of those self loathing 'jeets aren't ya. Did your Windows update delete the rest of your shitposting folder Ranjeet? Why are you always so upset, and nonsensical?
>p-p-pls believe me sir APLE IS LAG PLS BELIEVE MY WEEDIO
Why do Indians hate Apple so much? Everyone knows iDevices are silky smooth, why are you trying to troll?
Anthony King
>iPoojeet can't face reality >keeps damage controlling with every fake benchmark he can dig up except actual real world app performance which he pretends he didnt see COPE
>apple cpu close to desktop cpu So they finally started to support fuck ton of desktop specific instruction sets? And are not just a single core performance for 10 seconds until throttling? Nice I am throwing my i7 in the bin and purchasing iphone, I am sure autocad performance will increase at least 50%
Jayden Baker
>iShit >silky smooth How many times a day do you shit in the street Rajeet? Does Applel at least provide you with toilet paper?