Windows 10 Gaming Performance Update

[AKK_K]

Amd Ryzen Performance Update

ข้อมูลมาถึงวันนี้ชัดเจนขึ้นเป็นลำดับ
ถ้าเป็นไปตามนี้
ในแง่ประสิทธิภาพโดยรวมต่อราคา Intel แพ้ขาด
เพราะ แค่ RYZEN 5 1600X (6 core /12 thread) 259$
Cinebench R15 - Multithread ก็ เร็ว พอๆกับ I7 6800K (6 core /12 thread) 434$

Impressively, the 1600X which carries a similar price-tag to the unlocked Kaby Lake Core i5 7600K manages
to punch two hundred dollars above its weight and slightly outperofrm the much more expensive $430 i7 6800K.
This is despite running at 100MHz below the Intel Broadwell-E six core.
This suggests that the Zen microarchitecture may be a much more formidable IPC contender than we previously thought,
outperforming Broadwell core for core and clock for clock, at a lower wattage. Very impressive indeed.

http://wccftech.com/amd-ryzen-1600x-cinebench-r15-performance-confirmed/

และในแง่กลยุทธ Intel โดนหมัดชุดข้อมูลลงไปโดนนับอยู่ครับ ทำอะไรไม่ถูก งงเป็นไก่ตาแตก

ถ้ามีผลเทส เกม เช่น Watchdogs 2 (highly multi-threaded game) ออกมายืนยันอีกทีจะดีมากเลยครับ

[AKK_K]

once upon a time
กาลครั้งหนึ่งเมื่อไม่นานเท่าไร
2GB มันน่าจะได้ตายไปแล้ว

4GB ไปเลย 6GB
8GB ปรับสุด

ปรับสุดคุณได้อะไร #1.0

Assassin's Creed: Unity - GTX 1070 + i7 4790k (Ultra Settings) (MSAA x8) [1080p 60fps]
จั่วหัวคลิปว่า Msaa x8 60 fps แต่มันได้เท่าไรต้องดูเอง

Processador: Core i7 4790k 4.4ghz
Water Cooler Corsair H55
Memoria: 16GB RAM DDR3 1600mhz HyperX Fury
Placa de V?deo Primaria: Geforce GTX 1070 8GB G1 Gaming
SSD: PNY 120GB 2,5? SATA III 6GB/s
HD: SATA III 7200rpm 1TB
Fonte: Strike X 500w 80 plus bronze
OS: Windows 10 64bits

[AKK_K]

Blender ที่ใช้อ้างอิงว่า AMD RYZEN มีชัยเหนือ Intel

แต่ล่าสุดได้ออก Version 2.78b

และจากการทดสอบด้วย I5 4590
2.78b I5 4590 1 นาที 6 วิ
2.78a I5 4590 1 นาที 50 วิ

นั่นหมายความว่า Intel แรงขึ้นจากเดิม 68% (I5)

ตามคลิป Intel แรงขึ้น กว่า 40%

เกมจะพลิกไม้

blender https://www.blender.org/download/
data http://download.amd.com/demo/RyzenGraphic_27.blend

AMD Ryzen Blender Demo (2.78a VS Blender 2.78b time difference) I7-6850k @4.5GHz

[AKK_K]

แล้ว AMD RYZEN 8 core/16 thread จะชนะ I7 6700K หรือ 7700K @4.5 ไหม /Gaming

ตอนนี้คงไม่ชนะ

เพราะว่า

Haswell-E VS Skylake | i7 5960X VS 6700K 4.5GHz Gaming performance | GTX 1080 FE

[AKK_K]

และ
i7-6800K vs i7-6700K in 9 games (GTX 1070)

[AKK_K]

บทความจาก Eurogamer.net(Digital Foundry) มาพอดี อย่างกะนัดกันไว้

In Theory: How AMD's Ryzen will disrupt the gaming CPU market
By Richard Leadbetter Published 19/02/2017

To address gaming performance first, two variables are at play - IPC (instructions per clock) and frequency.
Having fast single-thread performance compared with high clock speeds tends to be vey important for gaming,
which explains why the Core i7 6700K managed to narrowly beat out both the six-core i7 5820K and the eight-core 5960X in our gaming CPU face-off last year.
Two generations of IPC improvements along with higher clocks - remarkably - proved to be far more beneficial than more cores.
Based on AMD's own tests, still to be confirmed outside of controlled conditions,
Ryzen has single-thread performance on par with Intel's older Broadwell architecture and the stated clocks are lower than the new i7 7700K.

http://www.eurogamer.net/articles/digitalfoundry-2017-how-amds-ryzen-will-disrupt-the-cpu-market

[AKK_K]

Intel AVX
Purpose

Intel Advanced Vector Extensions (AVX) is a 256-bit SIMD floating point vector extension of Intel architecture.
It includes extensions to both instruction and register sets.
Microsoft has developed some API enhancements, such as XState functions,
that enable applications to access and manipulate extended processor feature information and state, including Intel AVX.

//

Advanced Vector Extensions (AVX) are extensions to the x86 instruction set architecture for microprocessors from Intel and AMD
proposed by Intel in March 2008 and first supported by Intel with the Sandy Bridge[1] processor shipping in Q1 2011 and later on by AMD with the Bulldozer[2] processor shipping in Q3 2011. AVX provides new features, new instructions and a new coding scheme.


AVX2 expands most integer commands to 256 bits and introduces FMA. AVX-512 expands AVX to 512-bit support utilizing a new EVEX prefix encoding proposed by Intel in July 2013 and first supported by Intel with the Knights Landing processor scheduled to ship in 2015.

//

AVX2

AVX2 is an expansion of the new AVX instruction set introduced with Sandy Bridge.
The most important element of AVX was that Intel had improved the floating point units of the processor so they could process 256-bit numbers.
AVX came with 12 new instructions, some of which are suitable for three variables.
For example, it included an instructions for C = A + B. Prior to AVX, processor could only work with two variables, so the above equation had to go: A = A + B, and then C = A.

AVX2 takes this a step further. An important improvement in Haswell is that the integer execution units now can work with 256-bit numbers.
The number of instructions that can work with three variables has been increased, including instructions for multiplications and bit operations.
There are new instructions for retrieving data according to the gather-scatter method, which is important for vector calculations used by multimedia software.

[AKK_K]

Blender 2.78b Release Notes.

.
.
.

Cycles: CPU
Use more global SSE optimizations for SSE4.1+ kernels.
Multiple improvements for the latest AVX2 CPUS:
Optimized various math utilities (cross-products, dot-products, min-axis-selection and others).
Faster version of triangle intersection function.
Optimization of various steps in BVH traversal algorithm (including both construction and traversal).

.
.
.

[AKK_K]

AMD's core is actually 10 per cent smaller than Intel's.
The question is, are there any chinks in the armour? Can a company with an R&D budget a fraction of Intel's really compete?
Well, we do know that there are some compromises in the Ryzen design.
For example, Ryzen's support for AVX instructions is 128-bit in nature, meaning two cycles are required to achieve what an Intel core does in one

[AKK_K]

What common desktop applications are using AVX and AVX2?


Any application which uses Intel IPP / Intel MKL.
Any application which uses OpenBLAS, Apple Acceleration Library, Eigen.
Most of the Signal / Image Video Processing Filters (And Image and Video Compression libraries).
Compression utilities.
So we got something like:
Adobe Photoshop.
Adobe After Effects.
Adobe Premiere.
Photoshop / After Effects / Premiere Plug In's.
PeaZip / 7 Zip / WinRar.
x264 / FFMPEG (Used by many desktop applications, for instance, Firefox and VLC Player).
Excel.
MATLAB / Octave / Python (Numpy, Scipy, TensorFlow, etc...).
Julia.
Numeric Libraries (Used intensively in many applications).
I can go on and even be more specific.

It is much easier to vectorize code than to parallelize it.
Hence advancement like AVX & AVX2 are welcome.

Pay attention that what needed to support efficient vectorization is localization of the data.
Hence we need CPU's with larger and faster (Or same speed to the least) cache system.
We also need more bandwidth for the Main Memory system.

.
.
.

Don't know whether you consider PhysX a common desktop app, but it uses AVX for cloth simulation.
And if I were a betting man, I would say that NVidia would use AVX2 for the CPU version of Flex whenever it is released,
as it will use much more advanced physics effects like smoke and fluid simulation.

จาก

https://forums.anandtech.com/threads/what-common-desktop-applications-are-using-avx-and-avx2.2498660/

[AKK_K]

As I mentioned in my OP, PhysX 3.xx uses AVX for cloth solving, so it appears that SIMD is seeing a lot more use in the various physics engines than in the main game engine...

This comes directly from one of the programmers of PhysX, who was in fact responding to one of my own posts on Anandtech concerning CPU and GPU PhysX..

If anything, CPU Physics is on the up and up, as NVidia has started to throw a lot of resources into developing their PhysX 3.xx SDK, and it runs primarily on the CPU and is extremely fast.

Witcher 3 has some of the best cloth physics in a game, and it runs completely on the CPU. A lot of the effects seen in current PhysX 3.xx titles are even better than what was seen in GPU accelerated PhysX titles 5 years ago.

จาก https://forums.anandtech.com/threads...2460033/page-2

Only the cloth code uses AVX at the moment. There are separate AVX and non-AVX codepaths.

Nvidia Physx AVX AVX2

http://www.codercorner.com/blog/?p=1129

[AKK_K]

AVX2 Support in Visual Studio C++ Compiler

AVX2 is yet another extension to the venerable x86 line of processors, doubling the width of its SIMD vector registers to 256 bits, and adding dozens of new instructions.
AVX2 shipped with Intel?s latest processor micro-architecture, codenamed ?Haswell?. (Its official name is ?4th generation Intel? Core? processor family?).
As well as AVX2, Haswell supports other features to help make your code run faster: FMA (Fused Multiply Add) and BMI (Bit Manipulation Instructions), in particular.
Haswell chips appear in many of the latest PCs, laptops and tablets (including our own Surface Pro 2).

This extra silicon opens a new corner of the playing field for compiler-writers ? to take your C++ source and generate these new instructions,
making your code run faster than before. And so we are releasing our initial support for AVX2 with the CTP2 of Visual Studio Update.

How do you tell the VC++ compiler to generate AVX2 instruction? From the command line, include the /arch:AVX2 switch.
If you work within Visual Studio, the screenshot below shows how to set this option.(ดูต่อจาก link ข้างล่างนะครับ)

https://blogs.msdn.microsoft.com/vcblog/2014/02/28/avx2-support-in-visual-studio-c-compiler/

[AKK_K]

AVX-512

AVX-512 are 512-bit extensions to the 256-bit Advanced Vector Extensions SIMD instructions for x86 instruction set architecture (ISA) proposed by Intel in July 2013,
and is supported in Intel's Xeon Phi x200 (Knights Landing) processor.[1] AVX-512 is not the first 512-bit SIMD instruction set that Intel has introduced in processors.
The earlier 512-bit SIMD instructions used in Xeon Phi coprocessors, derived from Intel's Larrabee project, are similar but not binary compatible and only partially source compatible.[1]
AVX-512 consists of multiple extensions not all meant to be supported by all processors implementing them.
Only the core extension AVX-512F (AVX-512 Foundation) is required by all implementations.

[AKK_K]

AMD AVX

Each floating-point unit can accept 128-bit-wide instructions like AVX, and two units can fuse together to tackle 256-bit instructions such as AVX2.
Unfortunately, this does put Zen behind Intel's Haswell, Broadwell and Skylake architectures, which contain dedicated 256-bit-wide floating-point units;
a single dedicated unit will always be more efficient than two units fusing together to accomplish the same goal.
As such, in workloads that utilize 256-bit instructions, IPC will be somewhat closer to Ivy Bridge (although still likely ahead in terms of performance).
With that said, workloads of this type are a very limited scope, and for the most part, Zen will be very competitive with the aforementioned architectures.
Intel has hindered AVX2 support by refusing to incorporate the instructions into its lower-end products, which will also work in AMD's favor for Zen.

Zen's successor, known as Zen+, will bring dedicated 256-bit-wide floating-point units, as well as higher clock frequencies and other currently unknown features. However, Zen in 2017 needs to lay down the path for which its successors can build upon. This is predominantly the reasoning behind Zen's narrower compute units; it will be cheap to produce, and use less power for similar instructions. While it will struggle in very specific workloads as mentioned above, for most scenarios, that will encompass most consumers very well and its performance will be on-par, if not better than Intel's offerings in these instances.

[AKK_K]

ีเข้าใจ AVX เข้าใจ Intel
ผมก็เพิ่งจะรู้ในสิ่งที่ไม่รู้มาก่อน หรือว่าอาจจะมองผ่านมันไป
AVX.. จริงๆแล้วเป็นสิ่งที่จะถูกเรียกใช้มากขึ้นก็เพราะมันสามารถเพิ่มประสิทธิภาพการทำงานได้ 40-50% จนถึงหลักหลายร้อย %
ในขณะที่ IPC (perfomance/clock) เริ่มจะตัน อีกด้านหนึ่งคือ hardware ที่ออกแบบมาทำงานเฉพาะด้านเช่น AVX นี่แหละ
ที่สามารถเพิ่มประสิทธิภาพแบบก้าวกระโดด
แต่ . . .

AVX instructions are power hungry and hurt stability when overclocked

ไปอ่านต่อเองนะช่วงนี้
ตัดไปที่

http://www.anandtech.com/show/10969/the-intel-core-i57600k-91w-review-the-more-amenable-mainstream-performer/11