Overclock.net › Forums › Components › Memory › DDR2 native speed
New Posts  All Forums:Forum Nav:

DDR2 native speed

post #1 of 26
Thread Starter 
I can find no information that shows one should be setting the DDR2 dram to a higher ratio than 1:1 with the FSB.

I have been reading on this site that some people "think" that DDR2 should be run in the DDR2's dram speed and even if this is faster than the FSB it will then (supposedly) show bandwidth increases. I have tried this with:



and



The only difference being the ECC copy bit flip errors being corrected in the copy procedure due to a faster ram speed than FSB. NO bandwith increases otherwise.

Can anyone who believes otherwise point me in the direction of actual scientific proof from websites?

R
Roped In
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 2600k P8P67 Pro Rev (3.1) @ B3 XFX Radeon HD 6950 XXX 16GB G.SKILL Ripjaws 9-11-9-28 
Hard DriveOptical DriveOSMonitor
2 X Callisto Deluxe 25nm 60GB Plextor PX-B910SA 4x Blu-ray DVD-RW Win7 64 2 x Samsung 275T+ 
KeyboardPowerCaseMouse
Logitech G15 Corsair HX10000 Antec 1200 Logitech G7 
Mouse Pad
Splatter Game Pad 
  hide details  
Reply
Roped In
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 2600k P8P67 Pro Rev (3.1) @ B3 XFX Radeon HD 6950 XXX 16GB G.SKILL Ripjaws 9-11-9-28 
Hard DriveOptical DriveOSMonitor
2 X Callisto Deluxe 25nm 60GB Plextor PX-B910SA 4x Blu-ray DVD-RW Win7 64 2 x Samsung 275T+ 
KeyboardPowerCaseMouse
Logitech G15 Corsair HX10000 Antec 1200 Logitech G7 
Mouse Pad
Splatter Game Pad 
  hide details  
Reply
post #2 of 26
Ropey, I'm the guy you helped out on my question about that FSB:RAM ratio. I don't have any screenshot available now, but I can asure you that I have benchmarked both of the settings too (I used the memory benchmark from the Lavalys Everst Home Edition).

That's why I needed some more info, the 3:4 setting gave me a higher score on the benchmark for both reading and writing. And because I read here that running a higher RAM then FSB didn't do any good, I started doubting. That's why I asked for help.

If I think logical, since 3:4 is getting better results on the benchmark, I should think that it does increase performance when RAM (if DDR2) is running faster then FSB.

Edit: I will get you the screenshots tomorrow.
My System
(13 items)
 
  
CPUMotherboardOSMonitor
P4 650 Prescott 3.4GHz Asus P5GDC-V Deluxe i915G Windows XP PRO SP2 AG neovo F-417 LCD x 2 
KeyboardPowerCaseMouse
Cherry Q-TEC 400 Watt Custom Made Logitech Optical Mouse 
  hide details  
Reply
My System
(13 items)
 
  
CPUMotherboardOSMonitor
P4 650 Prescott 3.4GHz Asus P5GDC-V Deluxe i915G Windows XP PRO SP2 AG neovo F-417 LCD x 2 
KeyboardPowerCaseMouse
Cherry Q-TEC 400 Watt Custom Made Logitech Optical Mouse 
  hide details  
Reply
post #3 of 26
Thread Starter 
Quote:
Originally Posted by revmit
If I think logical, since 3:4 is getting better results on the benchmark, I should think that it does increase performance when RAM (if DDR2) is running faster then FSB.
You say logical but consider the physics. This is not addition of velocities at play here. If one stream is at 800MHz and another stream at 1000MHz then there is an increase of 200MHz streaming to the dram and a decrease of 200MHz when streaming to the CPU.

I fail to see how the physics can show an increase and no one has shown this to be the case. Algorithmic based benchmarks might well show what is not in reality there since it is simply numbers quantified on an algorithm.

So rather than saying an opinion, which we all have, I am looking for some form of scientific proof. Other than ram copy via ECC there is no showing of performance increases. Now there is no word out there on the net that what you say is the case so I still fall to the proven proof that so far there is not an addition of velocities just as there is not with DDR (and with DDR this has been factually proven).

I am not looking for screenshots as my screenshots show the opposite. I am looking for proof via a reputable scientific model testing. Anyone can post screenshots and they are NOT scientific proofs just as mine is not either.

R
Roped In
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 2600k P8P67 Pro Rev (3.1) @ B3 XFX Radeon HD 6950 XXX 16GB G.SKILL Ripjaws 9-11-9-28 
Hard DriveOptical DriveOSMonitor
2 X Callisto Deluxe 25nm 60GB Plextor PX-B910SA 4x Blu-ray DVD-RW Win7 64 2 x Samsung 275T+ 
KeyboardPowerCaseMouse
Logitech G15 Corsair HX10000 Antec 1200 Logitech G7 
Mouse Pad
Splatter Game Pad 
  hide details  
Reply
Roped In
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 2600k P8P67 Pro Rev (3.1) @ B3 XFX Radeon HD 6950 XXX 16GB G.SKILL Ripjaws 9-11-9-28 
Hard DriveOptical DriveOSMonitor
2 X Callisto Deluxe 25nm 60GB Plextor PX-B910SA 4x Blu-ray DVD-RW Win7 64 2 x Samsung 275T+ 
KeyboardPowerCaseMouse
Logitech G15 Corsair HX10000 Antec 1200 Logitech G7 
Mouse Pad
Splatter Game Pad 
  hide details  
Reply
post #4 of 26
No offense but that doesn't prove anything.

If we're going to talk about a "reputable" benchmark, it's rather not possible IMO.
As far as benchmarks go, they all have a flaw, hence the X850XTPE being pwned by an X1600 in 3Dmark06, you and I know which is the fastest, so much for "scientific proof".

Perhaps the thoguht of it hasn't sink in, but if there was no need for DDR2 Upto 1066 then why have nVidia and ATI gone from SD-RAM to DDR to DDR2 to GDDR3 (upto 800mhz) and isn't it ATI using GDDR4 now? I kinda see why? Because the raw power of the GPU needing it.

It's always been this way, when DDR came out, lots of people said they will never have the use for it.....2006 and everyone is close to make the transition to a full DDR2 system, yet DDR3 will be here soon @ 1333mhz speeds.

I've said it before and will say it again: I've seen the difference from 1:1 to 3:5 and 1:2 and will never look back at 1:1 again. The difference is not massive but I can see it while loading apps and games, not a with a large margin in difference but just noticable enough.
post #5 of 26
I don't know what to believe, I only know what I want to believe. And that is that DDR2 is different to SDRAM or DDR and can actually run faster then the FSB. So please keep on posting new facts about this issue, and I do mean facts... Because there's no convincing power in just plain argumenting your own thoughts.
My System
(13 items)
 
  
CPUMotherboardOSMonitor
P4 650 Prescott 3.4GHz Asus P5GDC-V Deluxe i915G Windows XP PRO SP2 AG neovo F-417 LCD x 2 
KeyboardPowerCaseMouse
Cherry Q-TEC 400 Watt Custom Made Logitech Optical Mouse 
  hide details  
Reply
My System
(13 items)
 
  
CPUMotherboardOSMonitor
P4 650 Prescott 3.4GHz Asus P5GDC-V Deluxe i915G Windows XP PRO SP2 AG neovo F-417 LCD x 2 
KeyboardPowerCaseMouse
Cherry Q-TEC 400 Watt Custom Made Logitech Optical Mouse 
  hide details  
Reply
post #6 of 26
Thread Starter 
Quote:
Originally Posted by revmit
I don't know what to believe
That's a mouthful.

Quote:
Originally Posted by revmit
please keep on posting new facts about this issue, and I do mean facts.
Please keep your requests of thread actions to your own created posts. This is simple forum etiquette.

Roped In
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 2600k P8P67 Pro Rev (3.1) @ B3 XFX Radeon HD 6950 XXX 16GB G.SKILL Ripjaws 9-11-9-28 
Hard DriveOptical DriveOSMonitor
2 X Callisto Deluxe 25nm 60GB Plextor PX-B910SA 4x Blu-ray DVD-RW Win7 64 2 x Samsung 275T+ 
KeyboardPowerCaseMouse
Logitech G15 Corsair HX10000 Antec 1200 Logitech G7 
Mouse Pad
Splatter Game Pad 
  hide details  
Reply
Roped In
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 2600k P8P67 Pro Rev (3.1) @ B3 XFX Radeon HD 6950 XXX 16GB G.SKILL Ripjaws 9-11-9-28 
Hard DriveOptical DriveOSMonitor
2 X Callisto Deluxe 25nm 60GB Plextor PX-B910SA 4x Blu-ray DVD-RW Win7 64 2 x Samsung 275T+ 
KeyboardPowerCaseMouse
Logitech G15 Corsair HX10000 Antec 1200 Logitech G7 
Mouse Pad
Splatter Game Pad 
  hide details  
Reply
post #7 of 26
Well my ram defaulted to 1:1 on my first boot, being 400mhz. I bumped it to 3:4 and now at 533 it runs much faster.
Main Gaming Rig
(14 items)
 
  
CPUMotherboardGraphicsRAM
2500K 4.8ghz Asus P8P67 Deluxe Sapphire 7970 1200/1600 8GB GSKILL SNIPER 
Hard DriveOptical DriveOSMonitor
128GB OCZ Vertex 4 LG 24X SATA Windows 7 Pro x64 Dell 22/24/22 LCDs 
KeyboardPowerCaseMouse
Logitech G110 Toughpower 1000watt 108A Corsair 800D Logitech G700 
Mouse PadAudio
SteelSeries QCK+ XFI XtremeMusic 
  hide details  
Reply
Main Gaming Rig
(14 items)
 
  
CPUMotherboardGraphicsRAM
2500K 4.8ghz Asus P8P67 Deluxe Sapphire 7970 1200/1600 8GB GSKILL SNIPER 
Hard DriveOptical DriveOSMonitor
128GB OCZ Vertex 4 LG 24X SATA Windows 7 Pro x64 Dell 22/24/22 LCDs 
KeyboardPowerCaseMouse
Logitech G110 Toughpower 1000watt 108A Corsair 800D Logitech G700 
Mouse PadAudio
SteelSeries QCK+ XFI XtremeMusic 
  hide details  
Reply
post #8 of 26
Thread Starter 
Yes, your ram runs much faster. However, that does not mean your cpu is running faster or that the data streaming from your cpu to your ram is running faster. Just that the ram itself is running faster.

CPU @ 800MHz fsb streams data to the FSB @ 800MHz and then to the RAM at 800MHz, once it is at the DRAM it is deposited at 1066 and then returned at 800MHz on the lines and to the 800MHz fsb processor.

But yes, whilst deposited in the DRAM it is being desposited at that 533. Put it this way. You are on a bus that is on the highway. The highway is 50KM/hr and the bus is moving at that speed. You are on the bus walking forward at 10KM/hr. This does not change the fact that the bus (fsb) is still running at 50KM/hr. Having five cars on a four lane highway does not make the fifth car any faster

What I mean is that the data has to drop to the speed of the fsb to transfer the data and only on the DRAM is it moving for copy/write at a speed faster than the fsb. Now if the data is being given to it as 800MHz then it can not pick up any faster than it is being given.

It's like having a small throated bottle of water turned upside down. The neck is the 800MHz fsb. No getting around that fact even if the body of the bottle is of larger capacity. If you fill the bottle at a certain speed then you are held to the size of the neck and if you pour the water out you are once again held to the size of the neck at that certain speed.

R
Roped In
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 2600k P8P67 Pro Rev (3.1) @ B3 XFX Radeon HD 6950 XXX 16GB G.SKILL Ripjaws 9-11-9-28 
Hard DriveOptical DriveOSMonitor
2 X Callisto Deluxe 25nm 60GB Plextor PX-B910SA 4x Blu-ray DVD-RW Win7 64 2 x Samsung 275T+ 
KeyboardPowerCaseMouse
Logitech G15 Corsair HX10000 Antec 1200 Logitech G7 
Mouse Pad
Splatter Game Pad 
  hide details  
Reply
Roped In
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 2600k P8P67 Pro Rev (3.1) @ B3 XFX Radeon HD 6950 XXX 16GB G.SKILL Ripjaws 9-11-9-28 
Hard DriveOptical DriveOSMonitor
2 X Callisto Deluxe 25nm 60GB Plextor PX-B910SA 4x Blu-ray DVD-RW Win7 64 2 x Samsung 275T+ 
KeyboardPowerCaseMouse
Logitech G15 Corsair HX10000 Antec 1200 Logitech G7 
Mouse Pad
Splatter Game Pad 
  hide details  
Reply
post #9 of 26
Ropey, I wasn't telling you or anyone what to post or not... I was thankfull of you guys trying to figure it out, because I would like some answers too. There's no need for you to be so offended.
My System
(13 items)
 
  
CPUMotherboardOSMonitor
P4 650 Prescott 3.4GHz Asus P5GDC-V Deluxe i915G Windows XP PRO SP2 AG neovo F-417 LCD x 2 
KeyboardPowerCaseMouse
Cherry Q-TEC 400 Watt Custom Made Logitech Optical Mouse 
  hide details  
Reply
My System
(13 items)
 
  
CPUMotherboardOSMonitor
P4 650 Prescott 3.4GHz Asus P5GDC-V Deluxe i915G Windows XP PRO SP2 AG neovo F-417 LCD x 2 
KeyboardPowerCaseMouse
Cherry Q-TEC 400 Watt Custom Made Logitech Optical Mouse 
  hide details  
Reply
post #10 of 26
Thread Starter 
LOL - - I'm not offended rev. I'm just anal
Roped In
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 2600k P8P67 Pro Rev (3.1) @ B3 XFX Radeon HD 6950 XXX 16GB G.SKILL Ripjaws 9-11-9-28 
Hard DriveOptical DriveOSMonitor
2 X Callisto Deluxe 25nm 60GB Plextor PX-B910SA 4x Blu-ray DVD-RW Win7 64 2 x Samsung 275T+ 
KeyboardPowerCaseMouse
Logitech G15 Corsair HX10000 Antec 1200 Logitech G7 
Mouse Pad
Splatter Game Pad 
  hide details  
Reply
Roped In
(13 items)
 
  
CPUMotherboardGraphicsRAM
Core i7 2600k P8P67 Pro Rev (3.1) @ B3 XFX Radeon HD 6950 XXX 16GB G.SKILL Ripjaws 9-11-9-28 
Hard DriveOptical DriveOSMonitor
2 X Callisto Deluxe 25nm 60GB Plextor PX-B910SA 4x Blu-ray DVD-RW Win7 64 2 x Samsung 275T+ 
KeyboardPowerCaseMouse
Logitech G15 Corsair HX10000 Antec 1200 Logitech G7 
Mouse Pad
Splatter Game Pad 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Memory
Overclock.net › Forums › Components › Memory › DDR2 native speed