News
 

Bookmark and Share

(9) 

Intel Corp., the world's largest maker of microprocessors and supporting logic, will reveal loads of previously unheard details about its next-generation code-named Haswell micro-architecture at the Intel Developer Forum next month. The chipmaker will discuss all the peculiarities of the new chip design and will likely even demonstrate it in action.

During four sessions dedicated to Haswell, Intel will discuss the new processor innovations in general as well as advanced vector extensions 2 (AVX 2), bit manipulation new instructions (BMI) as well as transactional synchronization extensions (TSX) instructions and their benefits in particular. In addition, the company will have a session called "Experiencing the 3rd and 4th Generation Intel Core Microarchitecture", which implies that the company will compare both current Ivy Bridge as well as future Haswell capabilities.

 

Since the world's largesty chipmaker has demonstrated Haswell chips for a number of times already, this time the company will likely show it working. Unfortunately, it is unlikely that the company will discuss performance benefits of the new micro-architecture in details.

Intel Haswell microprocessors for mainstream desktops and laptops will be structurally similar to existing Core i-series "Sandy Bridge" and "Ivy Bridge" chips and will continue to have two or four cores with Hyper-Threading technology along with graphics adapter that shares last level cache (LLC) with processing cores and works with memory controller via system agent. The processors that belong to the Haswell generation will continue to rely on dual-channel DDR3/DDR3L  memory controller with DDR power gating support to trim idle power consumption. The chip will have three PCI Express 3.0 controllers, Intel Turbo Boost technology with further improvements and so on.

On the micro-architectural level the chip will be a lot dissimilar compared to available solutions. It is believed that the new Haswell x86 micro-architecture will be substantially different from existing, which will enable further scalability and performance increases. Besides, Haswell will support numerous new instructions, including AVX2,  bit manipulation instructions, FPMA (floating point multiple accumulate) and others. The new graphics core based on Denlow architecture is projected to support such new features as DirectX 11.1, OpenGL 3.2+, to be substantially more powerful and to be certified to run numerous professional applications.

Intel Haswell chips also implement a number of aggressive measures to trim power consumption, including power aware interrupt routing for power/performance optimizations, configurable TDP and LPM, DDR power gating, power optimizer (CPPM) support, idle power improvements, latest power states, etc.

Tags: Intel, Ivy Bridge, Haswell, Denlow, 22nm, DirectX, OpenGL, TSX

Discussion

Comments currently: 9
Discussion started: 08/10/12 12:24:43 AM
Latest comment: 09/04/12 11:46:28 AM
Expand all threads | Collapse all threads

[1-4]

1. 
"graphics adapter that shares last level cache (LLC) with processing cores"

Incorrect. The graphics adapter will have dedicated on-package memory, and lots of it...and I it's graphics-use-only. This is the secret sauce that will destroy AMD's APU scheme.
0 1 [Posted by: AnonymousGuy  | Date: 08/10/12 12:24:43 AM]
Reply
- collapse thread

 
That's not what I heard, I've read by many different sources that in order to allow things like OpenCL to work better on the CPU and iGPU that they will share this extra LLC (Last Level Cache) or L4 Cache. This is because one of the main bottle necks with OpenCL is the copy operation. Also the Chip is to be two chips in one, as the 64 or 128 MB of L4 is going to be applied after the fact. This should increase yields as a small defect in the L4 would render the whole chip worthless. SkyLark is meant to go further with this idea and allow you to run code within the iGPU or the CPU, blurring the idea of what a conventional CPU should do.
0 1 [Posted by: Dygear  | Date: 08/10/12 04:58:05 AM]
Reply
 
LOL, yeah... This generation Intel is finally going to smoke AMD and Nvidia in graphics, right...

With optimizations such as:

* Incorrect OpenGL rendering
* Cache
* Maybe stacked memory(or not)
* Intel GPU silicon awesomeness
* Intel GPU driver excellence

Of course, real GPUs don't have much if any cache, and rely on ridiculous memory bandwidth, because that's what they actually need, not hacks like 2MB L6 cache.

If a powerful GPU needs 100GB/s bandwidth to stay busy, then cache isn't going to help much with 20GB/s memory bandwidth or whatever DDR3 will be doing. You act like cache is some revolutionary idea that AMD just hasn't thought of yet. The GPU experts at AMD and Nvidia just know that stream processors are a better use of GPU silicon than cache for overall performance, where Intel has to use every hack they can to create the illusion that their GPU isn't so bad.

Furthermore, since Intel is once again comparing Haswell's top GPU to Ivy Bridge's entry-level GPU in their slides to get that 3x faster figure, then we know for sure that they still suck, because the top-to-top comparison once again fails to impress.
2 2 [Posted by: h1719118  | Date: 08/10/12 09:44:14 AM]
Reply
 
"that AMD just hasn't thought of yet"

AMD doesn't have the process technology to do what Intel is going to do with Haswell's gfx. It's lot's more real memory right next to the die. And 3x faster is the performance gain with the increased graphic units alone. The dedicated memory is going to take those gains even higher.

Mark my words, Haswell will leapfrog Trinity. Once that happens, game over for AMD. Because at that point 50 fps vs 60 fps isn't going to matter, it's all good enough for integrated graphics.
2 1 [Posted by: AnonymousGuy  | Date: 08/10/12 06:56:51 PM]
Reply
 
show the post
0 3 [Posted by: keysplayer  | Date: 08/11/12 03:53:34 PM]
Reply

2. 
Will be interesting to see, how much perf gain from transactional memory support
0 0 [Posted by: Tristan  | Date: 08/10/12 03:10:27 AM]
Reply

3. 
Intel paid Nvidia 1.5 billion for Technology Licensing Fees (Patents) for its IGP, So its not as if Intel don’t have the tech or patents, it’s the knowhow. AMD has an advantage here slightly in the IGP as do Intel in the CPU.

http://pressroom.nvidia.c...mp;releasejsp=release_157

Nvidia deserves everything it gets including in the HPC market, Its also why you see non existent or useless nvidia products in the bottom end bcoz it doesn't have a clue, whats coming from AMD or Intel

http://arstechnica.com/bu.../?comments=1#comments-bar
0 1 [Posted by: keysplayer  | Date: 08/11/12 04:17:49 PM]
Reply

4. 
This is a waste of time by Intel it should be or it may be too late to think about producing CPU's for the coming tech. not sticking to old tech with x86; as ARM type CPU's are coming as it is now ARM CPU's outsell PC type chips,(Even the graphic part of their newer designs is surpassed by the ARM family. Remember when one becomes a Near monopoly there is only one direction to go.) etc by a long way. Microsoft has seen the writing on the wall and so has AMD.
1 0 [Posted by: tedstoy  | Date: 08/12/12 05:10:45 PM]
Reply

[1-4]

Add your Comment




Related news

Latest News

Wednesday, October 8, 2014

8:52 pm | Lisa Su Appointed as New CEO of Advanced Micro Devices. Rory Read Steps Down, Lisa Su Becomes New CEO of AMD

Thursday, August 28, 2014

12:22 pm | AMD Has No Plans to Reconsider Recommended Prices of Radeon R9 Graphics Cards. AMD Will Not Lower Recommended Prices of Radeon R9 Graphics Solutions

Wednesday, August 27, 2014

9:09 pm | Samsung Begins to Produce 2.13GHz 64GB DDR4 Memory Modules. Samsung Uses TSV DRAMs for 64GB DDR4 RDIMMs

Tuesday, August 26, 2014

6:41 pm | AMD Quietly Reveals Third Iteration of GCN Architecture with Tonga GPU. AMD Unleashes Radeon R9 285 Graphics Cards, Tonga GPU, GCN 1.2 Architecture

Monday, August 25, 2014

6:05 pm | Chinese Inspur to Sell Mission-Critical Servers with AMD Software, Power 8 Processors. IBM to Enter Chinese Big Data Market with the Help from Inspur