- Joined
- Nov 10, 2006
- Messages
- 4,667 (0.68/day)
- Location
- Washington, US
| System Name | Rainbow |
|---|---|
| Processor | Intel Core i7 8700k |
| Motherboard | MSI MPG Z390M GAMING EDGE AC |
| Cooling | Corsair H115i, 2x Noctua NF-A14 industrialPPC-3000 PWM |
| Memory | G. Skill TridentZ RGB 4x8GB (F4-3600C16Q-32GTZR) |
| Video Card(s) | ZOTAC GeForce RTX 3090 Trinity |
| Storage | 2x Samsung 950 Pro 256GB | 2xHGST Deskstar 4TB 7.2K |
| Display(s) | Samsung C27HG70 |
| Case | Xigmatek Aquila |
| Power Supply | Seasonic 760W SS-760XP |
| Mouse | Razer Deathadder 2013 |
| Keyboard | Corsair Vengeance K95 |
| Software | Windows 10 Pro |
| Benchmark Scores | 4 trillion points in GmailMark, over 144 FPS 2K Facebook Scrolling (Extreme Quality preset) |
In reply to Common LGA-1366 Cooling Myths Busted. I moved it here because my post became long and rather off the original topic.
Wanted to say thanks to candle_86 and everyone else who put the time in to writing an educational post. Lets see if I can add a little more. (I hope it's all right.
)
No one mentioned AMD's role in the rise of mainstream 64-bit processing.
Intel designed the Itanium architecture early on. It was designed for servers, it was designed for 64-bit computing, it was designed to handle more than one instruction per cycle (via EPIC). The big problem was that it was not backwards compatible with 32-bit applications. In fact, it wasn't even remotely affiliated with x86 at all. This was the very new, very different world of IA-64. Upgrading to an Itanium server meant ditching all your old server software, buying everything new and starting from scratch.
When 32-bit servers started to come closer to their memory addressing limit, AMD saw a demand for 64-bit memory addressing. They knew, thanks to Intel's flop with Itanium, that people simply would not dive in to a new architecture head first. Like the move from 16-bit to 32-bit, the consumers wanted backwards compatibility. With that in mind, AMD designed and nicely executed their 64-bit extension to 32-bit processors. x86-64, later renamed to AMD64. By making 64-bit processing an extension of the original x86 architecture, backwards compatibility was not only retained, 32-bit performance was downright identical (if not better!). Intel quickly cloned this extended instruction set under the name EM64T. Itanium would never touch the desktop.
The moral of the story is that AMD has brought some major technology to the processor market. Integrated Memory Controllers, HyperTransport and x86-64 to name a few. Intel has cloned all three.
AMD wasn't the only people who designed HyperTransport and it is in no way an AMD-only technology.
As video cards and other quickly growing I/O speeds (SATA, USB2.0, etc) demanded more throughput, they pushed conventional FSB to it's transfer limit. I may get this wrong, but a 64-bit 133MHz bus DDR does (133MHz * 64-bit * DDR / 8 * 2-way =) 2,128 MB/s. PC133 was takes up to 1,064 MB/s. AGP 8x demanded its own 2133 MB/s. Don't forget everything the southbridge wants.. This led to components struggling to share the bandwidth.
Intel's introduced "Quad Pumping" which bought them more time. This effectively quadrupled device bandwidth allowing for more breathing room between devices. AMD followed shortly after, introducing HyperTransport to their processors.
HyperTransport was designed by the HyperTransport Consortium (Advanced Micro Devices, Alliance Semiconductor, Apple Computer, Broadcom Corporation, Cisco Systems, NVIDIA, PMC-Sierra, Sun Microsystems, and Transmeta).
Today's revision of HyperTransport currently can transfer up to 52 GB/s across it's bus, supporting backwards compatibility to older revisions.
Now, instead of simply licensing HyperTransport, Intel decided they would rather design their own interconnect (apparently now dubbed Intel QuickPath Interconnect) which is said to show up in Nehalem.
So, yes. Intel makes advancements on things like SSE, but AMD is making enhancements on things like Integrated Memory Controllers. As candle_86 said, everyone shares.
I personally think Intel should just hop on with HyperTransport. It's tried and true. AMD already worked the bugs out for them. They're copying the IMC and HT (and rumors of copying Fusion).. Why not just go for HyperTranport too?
Wanted to say thanks to candle_86 and everyone else who put the time in to writing an educational post. Lets see if I can add a little more. (I hope it's all right.
)No one mentioned AMD's role in the rise of mainstream 64-bit processing.
Intel designed the Itanium architecture early on. It was designed for servers, it was designed for 64-bit computing, it was designed to handle more than one instruction per cycle (via EPIC). The big problem was that it was not backwards compatible with 32-bit applications. In fact, it wasn't even remotely affiliated with x86 at all. This was the very new, very different world of IA-64. Upgrading to an Itanium server meant ditching all your old server software, buying everything new and starting from scratch.
When 32-bit servers started to come closer to their memory addressing limit, AMD saw a demand for 64-bit memory addressing. They knew, thanks to Intel's flop with Itanium, that people simply would not dive in to a new architecture head first. Like the move from 16-bit to 32-bit, the consumers wanted backwards compatibility. With that in mind, AMD designed and nicely executed their 64-bit extension to 32-bit processors. x86-64, later renamed to AMD64. By making 64-bit processing an extension of the original x86 architecture, backwards compatibility was not only retained, 32-bit performance was downright identical (if not better!). Intel quickly cloned this extended instruction set under the name EM64T. Itanium would never touch the desktop.
The moral of the story is that AMD has brought some major technology to the processor market. Integrated Memory Controllers, HyperTransport and x86-64 to name a few. Intel has cloned all three.
AMD wasn't the only people who designed HyperTransport and it is in no way an AMD-only technology.
As video cards and other quickly growing I/O speeds (SATA, USB2.0, etc) demanded more throughput, they pushed conventional FSB to it's transfer limit. I may get this wrong, but a 64-bit 133MHz bus DDR does (133MHz * 64-bit * DDR / 8 * 2-way =) 2,128 MB/s. PC133 was takes up to 1,064 MB/s. AGP 8x demanded its own 2133 MB/s. Don't forget everything the southbridge wants.. This led to components struggling to share the bandwidth.
Intel's introduced "Quad Pumping" which bought them more time. This effectively quadrupled device bandwidth allowing for more breathing room between devices. AMD followed shortly after, introducing HyperTransport to their processors.
HyperTransport was designed by the HyperTransport Consortium (Advanced Micro Devices, Alliance Semiconductor, Apple Computer, Broadcom Corporation, Cisco Systems, NVIDIA, PMC-Sierra, Sun Microsystems, and Transmeta).
Today's revision of HyperTransport currently can transfer up to 52 GB/s across it's bus, supporting backwards compatibility to older revisions.
Now, instead of simply licensing HyperTransport, Intel decided they would rather design their own interconnect (apparently now dubbed Intel QuickPath Interconnect) which is said to show up in Nehalem.
So, yes. Intel makes advancements on things like SSE, but AMD is making enhancements on things like Integrated Memory Controllers. As candle_86 said, everyone shares.
I personally think Intel should just hop on with HyperTransport. It's tried and true. AMD already worked the bugs out for them. They're copying the IMC and HT (and rumors of copying Fusion).. Why not just go for HyperTranport too?
Last edited:
![[I.R.A]_FBi](https://tpucdn.com/forums/data/avatars/l/39/39344.jpg?1384761935){kind=avatar}
AMD.
