Overclock.net › Forums › Industry News › Hardware News › [TGD] How a forgotten Intel invention could revolutionize the CPU
New Posts  All Forums:Forum Nav:

[TGD] How a forgotten Intel invention could revolutionize the CPU - Page 4

post #31 of 45
Quote:
Originally Posted by Danker16 View Post
they get power and they want to F*** things up..what an advantage... GO AMD!
they aren't allowed to make a discovery? what do you want to do, lock every intel research employee in jail? do you realize how ridiculous your statement is
p
(13 items)
 
  
CPUMotherboardGraphicsRAM
Athlon X2 5000+ BE @ 3.0GHz BIOSTAR TForce TF570SLI 2 X PNY 9800GX2 1GB 512-bit 8GB OCZ Gold DDR2 800 
Hard DriveOptical DriveOSMonitor
36GB Raptor X 4 in RAID10 20x DVD+RW Slackware 12.1 Dual Samsung 24" WS LCD 
KeyboardPowerCaseMouse
IBM Model M PC P&C 610 Silencer (12V@49A) Ultra Aluminus Full Tower Logitech G5 
Mouse Pad
Ratpadz GS 
  hide details  
Reply
p
(13 items)
 
  
CPUMotherboardGraphicsRAM
Athlon X2 5000+ BE @ 3.0GHz BIOSTAR TForce TF570SLI 2 X PNY 9800GX2 1GB 512-bit 8GB OCZ Gold DDR2 800 
Hard DriveOptical DriveOSMonitor
36GB Raptor X 4 in RAID10 20x DVD+RW Slackware 12.1 Dual Samsung 24" WS LCD 
KeyboardPowerCaseMouse
IBM Model M PC P&C 610 Silencer (12V@49A) Ultra Aluminus Full Tower Logitech G5 
Mouse Pad
Ratpadz GS 
  hide details  
Reply
post #32 of 45
Quote:
Originally Posted by Licht View Post
Virtually no latency? Think again man. The L3 cache has like 4ns latency right there. This would have a couple hundred if not a few ms.
It would be faster than latency on standard DRAM, and if they can fit enough of it in there, your system speeds would be very fast.
There is also nothing that says that you couldn't still have standard DRAM in the system, this could be in addition to that.
    
CPUMotherboardGraphicsRAM
Q9650 E0 @ 4GHz (445x9) on 1.325V (bios) | ASUS Rampage Formula x48 | | BIOS 0410 | Visiontek 4870 790/1100 | 2x2GB G.Skill PI @1069MHz | 1.9V | 5-5-5-15 2T | 
Hard DriveOptical DriveOSMonitor
Seagate 2x320GB 7200.10 in RAID 0 + 250GB 7200.10 Lite-On DVD+-R/RW DL Vista Ultimate x64 42" Aquous @ 1440 x 900 (1920 x 1080 is too small) 
KeyboardPowerCaseMouse
Logitech MX5500 Silverstone DA750 THE Rocketfish Logitech MX Revolution 
Mouse Pad
Neat one from Body Worlds exhibit 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Q9650 E0 @ 4GHz (445x9) on 1.325V (bios) | ASUS Rampage Formula x48 | | BIOS 0410 | Visiontek 4870 790/1100 | 2x2GB G.Skill PI @1069MHz | 1.9V | 5-5-5-15 2T | 
Hard DriveOptical DriveOSMonitor
Seagate 2x320GB 7200.10 in RAID 0 + 250GB 7200.10 Lite-On DVD+-R/RW DL Vista Ultimate x64 42" Aquous @ 1440 x 900 (1920 x 1080 is too small) 
KeyboardPowerCaseMouse
Logitech MX5500 Silverstone DA750 THE Rocketfish Logitech MX Revolution 
Mouse Pad
Neat one from Body Worlds exhibit 
  hide details  
Reply
post #33 of 45
Quote:
Originally Posted by RPIJG View Post
It would be faster than latency on standard DRAM, and if they can fit enough of it in there, your system speeds would be very fast.
There is also nothing that says that you couldn't still have standard DRAM in the system, this could be in addition to that.
I was just pointing out that not even an L cache has no latency. Its important to realize this. (Although yeah the latency would be massively reduced, great stuff but you get my point now I'm sure.)
Lee XT
(17 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX-6300 Asus M5A97 SAPPHIRE Radeon HD 7850 AMD 4GB DDR3 1333MHZ 
RAMRAMRAMHard Drive
AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ OCZ Vertex 4 256GB 
CoolingOSMonitorKeyboard
Corsair H80 Windows 8.1 Pro MCE Dell P2414H WHXV7  Microsoft Generic 
PowerCaseMouseMouse Pad
Ultra 600W Limited Edition NZXT Black Steel Razer Deathadder Razer Goliath 
Audio
Realtek HD Audio 
  hide details  
Reply
Lee XT
(17 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX-6300 Asus M5A97 SAPPHIRE Radeon HD 7850 AMD 4GB DDR3 1333MHZ 
RAMRAMRAMHard Drive
AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ OCZ Vertex 4 256GB 
CoolingOSMonitorKeyboard
Corsair H80 Windows 8.1 Pro MCE Dell P2414H WHXV7  Microsoft Generic 
PowerCaseMouseMouse Pad
Ultra 600W Limited Edition NZXT Black Steel Razer Deathadder Razer Goliath 
Audio
Realtek HD Audio 
  hide details  
Reply
post #34 of 45
1. People don't want Intel to introduce either DRAM L cache on CPU or DRAM-on-die because they are afraid of change
2. Remember how physics cards were external(it's own card ) well now it's adopted into the chip itself, before that there were audio cards, now it's integrated into mobos, also network controllers and cards, you get the drift. So integration is the way of advancement, if you don't like it go use a 486 or P-I.
3. There are two directions Intel is likely to go in the future, 1. optical processing, 2. further integration by enlarging the die to double-tripple to fit DRAM+miscellenious improvements that benefit from integration. Intel could also follow #2, and then when optical processing matures do #1.

Note: Integration of 4-16 CPUs, plus graphic processors, plus DRAM = a lot of heat, but because the die is now 2-3 times bigger, the cooling becomes easier with greater surface area. Remember, they shrink the size in order to speed up latency of wires\\paths in the processor and reduce heat. The overall speed and latency is many times smaller inside the integrated CPU/Graphics/DRAM and 2-3 times bigger in size compared to external DRAM, video card and else, the value of integration is evident.
Edited by dumont - 6/19/08 at 4:08pm
GONDOR
(13 items)
 
  
CPUMotherboardGraphicsRAM
P4 2.8GHz @ 800FSB Asus P4C800-E Deluxe Asus GeForce FX 5900 Ultra 256MB OCZ 1GB (512MBx2) 
Hard DriveOptical DriveOSMonitor
Too Many Memorex (Prefer Lite-On) Classified Dell 2405WFP 
KeyboardPowerCaseMouse
Microsoft Internet White Vantec 420W Lian-Li PC601 Microsoft IntelliMouse Explorer 3.0 
Mouse Pad
Classified 
  hide details  
Reply
GONDOR
(13 items)
 
  
CPUMotherboardGraphicsRAM
P4 2.8GHz @ 800FSB Asus P4C800-E Deluxe Asus GeForce FX 5900 Ultra 256MB OCZ 1GB (512MBx2) 
Hard DriveOptical DriveOSMonitor
Too Many Memorex (Prefer Lite-On) Classified Dell 2405WFP 
KeyboardPowerCaseMouse
Microsoft Internet White Vantec 420W Lian-Li PC601 Microsoft IntelliMouse Explorer 3.0 
Mouse Pad
Classified 
  hide details  
Reply
post #35 of 45
Quote:
Originally Posted by Licht View Post
Virtually no latency? Think again man. The L3 cache has like 4ns latency right there. This would have a couple hundred if not a few ms.
I would think that properly done, this 2T DRAM might be able to achieve super low latencies, compared to regular DRAM and even SRAM. Since the physical distance from the DRAM to the MCT is dramatically shortened, that would remove a chunk of the latency. Next, since the size of the 2T DRAM cell is dramatically smaller than a SRAM cell, the overall array of 2T DRAM cells is smaller ie. the silicon area of equal data storage would be smaller for the 2T DRAM than SRAM. Because of this net decrease in size, the latency will be further reduced. If this is really able to scale to core clocks, the clockspeed would be much higher reducing even more latency. And finally, if this can outperform L2 cache, then there a possibility that L2 cache might be fully obsolete and omitted from the CPU design. With no L2 cache, there is one less stop or one less possible cache miss for data, reducing the overall latency.

Also, don't forget the other implications as well. Using only a third of the silicon real estate than SRAM, DRAM would lead to smaller dies, making each die more cost effective or at least it could free up real silicon real estate for more logic transistors. It would also be more power efficient as well and in this day in age, who doesn't want to save power and/or save the environment!

The sad thing is that AMD has a liscense for ZRAM, which is even twice as dense as this 2T DRAM, but AMD hasn't mentioned anything about the progress of the development. The last thing mentioned was that it wasn't scaling to speeds that was beneficial. Hopefully that will all change with high-K 45nm SOI.......
    
CPUMotherboardGraphicsRAM
X2 5000+ BE Gigabyte GA-MA78GM-S2H AM2+ Integrated HD3200 2x2GB OCZ DDR2 
Hard DriveMonitorPowerCase
Maxtor 40GB Samsung 2220WM 22" Antec Neo HE 550W Ultra Wizard 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
X2 5000+ BE Gigabyte GA-MA78GM-S2H AM2+ Integrated HD3200 2x2GB OCZ DDR2 
Hard DriveMonitorPowerCase
Maxtor 40GB Samsung 2220WM 22" Antec Neo HE 550W Ultra Wizard 
  hide details  
Reply
post #36 of 45
i can't wait till quantum computing comes mainstream, 1+ T(era)b/s or 1+ P(eta)b/s.......
post #37 of 45
I doubt quantum computing will be anywhere near available near mainstream within the next couple of centuries primarily for the reasons of physics, materials, and general cost manufacture and maintenance. Therefore it will be just cheaper to have high powered, multicore, relatively cool compared to today's cpus optical processors. It will be cheaper, common materials, cool temps, scalability, and easier to maintain (also have a longer operating cycle).
GONDOR
(13 items)
 
  
CPUMotherboardGraphicsRAM
P4 2.8GHz @ 800FSB Asus P4C800-E Deluxe Asus GeForce FX 5900 Ultra 256MB OCZ 1GB (512MBx2) 
Hard DriveOptical DriveOSMonitor
Too Many Memorex (Prefer Lite-On) Classified Dell 2405WFP 
KeyboardPowerCaseMouse
Microsoft Internet White Vantec 420W Lian-Li PC601 Microsoft IntelliMouse Explorer 3.0 
Mouse Pad
Classified 
  hide details  
Reply
GONDOR
(13 items)
 
  
CPUMotherboardGraphicsRAM
P4 2.8GHz @ 800FSB Asus P4C800-E Deluxe Asus GeForce FX 5900 Ultra 256MB OCZ 1GB (512MBx2) 
Hard DriveOptical DriveOSMonitor
Too Many Memorex (Prefer Lite-On) Classified Dell 2405WFP 
KeyboardPowerCaseMouse
Microsoft Internet White Vantec 420W Lian-Li PC601 Microsoft IntelliMouse Explorer 3.0 
Mouse Pad
Classified 
  hide details  
Reply
post #38 of 45
Quote:
Originally Posted by dumont View Post
I doubt quantum computing will be anywhere near available near mainstream within the next couple of centuries primarily for the reasons of physics, materials, and general cost manufacture and maintenance. Therefore it will be just cheaper to have high powered, multicore, relatively cool compared to today's cpus optical processors. It will be cheaper, common materials, cool temps, scalability, and easier to maintain (also have a longer operating cycle).
Thats what they said when vacuum tubes based computers were taking up entire buildings and requiring AC just to keep from catching on fire. Then the microprocessor came out based on similar ideals, revolutionized computing.
Lee XT
(17 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX-6300 Asus M5A97 SAPPHIRE Radeon HD 7850 AMD 4GB DDR3 1333MHZ 
RAMRAMRAMHard Drive
AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ OCZ Vertex 4 256GB 
CoolingOSMonitorKeyboard
Corsair H80 Windows 8.1 Pro MCE Dell P2414H WHXV7  Microsoft Generic 
PowerCaseMouseMouse Pad
Ultra 600W Limited Edition NZXT Black Steel Razer Deathadder Razer Goliath 
Audio
Realtek HD Audio 
  hide details  
Reply
Lee XT
(17 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX-6300 Asus M5A97 SAPPHIRE Radeon HD 7850 AMD 4GB DDR3 1333MHZ 
RAMRAMRAMHard Drive
AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ AMD 4GB DDR3 1333MHZ OCZ Vertex 4 256GB 
CoolingOSMonitorKeyboard
Corsair H80 Windows 8.1 Pro MCE Dell P2414H WHXV7  Microsoft Generic 
PowerCaseMouseMouse Pad
Ultra 600W Limited Edition NZXT Black Steel Razer Deathadder Razer Goliath 
Audio
Realtek HD Audio 
  hide details  
Reply
post #39 of 45
Quantum computing is ill suited to speed up 99% of computing tasks. It is good at solving things such as the Traveling Salesman Problem or factoring large numbers (important for cryptography), but for normal arithmetic calculation you're better off with a CPU.
It goes to eleven
(13 items)
 
  
CPUMotherboardGraphicsRAM
E6300 DS3 EVGA 8600GTS 2GB XMS2 DDR2-800 
Hard DriveOSMonitorKeyboard
1.294 TB Arch Linux/XP Samsung 226bw Eclipse II 
PowerCaseMouse
Corsair 520HX Lian-Li v1000B Plus G7 
  hide details  
Reply
It goes to eleven
(13 items)
 
  
CPUMotherboardGraphicsRAM
E6300 DS3 EVGA 8600GTS 2GB XMS2 DDR2-800 
Hard DriveOSMonitorKeyboard
1.294 TB Arch Linux/XP Samsung 226bw Eclipse II 
PowerCaseMouse
Corsair 520HX Lian-Li v1000B Plus G7 
  hide details  
Reply
post #40 of 45
I'm not quick on analogies but comparing oranges and apples only proves a lack of information about the subject.

There are concepts and there are physical limitations, for quantum computing the physical limitations are clear, at least with the molecular chemistry table we have at the moment and it's not like the table or the materials set will change enough to allow for what you are suggesting couple with physical limitations of quantum processing.
Edited by dumont - 6/22/08 at 6:36pm
GONDOR
(13 items)
 
  
CPUMotherboardGraphicsRAM
P4 2.8GHz @ 800FSB Asus P4C800-E Deluxe Asus GeForce FX 5900 Ultra 256MB OCZ 1GB (512MBx2) 
Hard DriveOptical DriveOSMonitor
Too Many Memorex (Prefer Lite-On) Classified Dell 2405WFP 
KeyboardPowerCaseMouse
Microsoft Internet White Vantec 420W Lian-Li PC601 Microsoft IntelliMouse Explorer 3.0 
Mouse Pad
Classified 
  hide details  
Reply
GONDOR
(13 items)
 
  
CPUMotherboardGraphicsRAM
P4 2.8GHz @ 800FSB Asus P4C800-E Deluxe Asus GeForce FX 5900 Ultra 256MB OCZ 1GB (512MBx2) 
Hard DriveOptical DriveOSMonitor
Too Many Memorex (Prefer Lite-On) Classified Dell 2405WFP 
KeyboardPowerCaseMouse
Microsoft Internet White Vantec 420W Lian-Li PC601 Microsoft IntelliMouse Explorer 3.0 
Mouse Pad
Classified 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Hardware News
Overclock.net › Forums › Industry News › Hardware News › [TGD] How a forgotten Intel invention could revolutionize the CPU