Overclock.net › Forums › Intel › Intel CPUs › FSB and Dividers: My Findings
New Posts  All Forums:Forum Nav:

FSB and Dividers: My Findings - Page 4

post #31 of 85
Quote:
Originally Posted by MrBungle View Post
#1 bothers me because you're lowering the multi so much. i don't think your motherboard will like that. i don't like #2 because that's the only setup where you're actually underclocking the memory. with #3 i don't know if your memory will be able to handle it. i think #4 looks pretty solid. i'm a pretty conservative overclocker though.

another option is to just set everything to 1337. it won't work, but it will look really cool in your BIOS

Wow. I have to give you some credit, your pretty good with this stuff. #1 doesn't run on my motherboard, you were right, for some odd reason it craps out at 1473FSB. (I can run 1500+ with 6x multi, weird eh?)

#2 is where I am underclocking my mem so I thought the same thing.

#3 Is most likely what I will use, because my memory can easily handle it

#4 is what I would use if my memory couldn't take it.

But now I found something else. Check this out:

FSB 1100
Multi 8x
Unlinked Ratio: 11:16
Memory 800mhz

These are direct from CPU-Z. All I wanted to you see here is what is in bold.
I noticed how on unlinked I still got a ratio in CPU-Z.

Does this mean that unlinked is still linked, just you can set your own ratio so to speak?

If so, then the tests that choggs is running will be true!!!
(see, we haven't totally hijacked his thread yet )
Smilodon
(13 items)
 
  
CPUMotherboardGraphicsRAM
e4400 @ 2.9Ghz EVGA 650i Ultra ATX MSI 8800GTS 640 2x1GB G.Skill DDR2 800 
Hard DriveOSMonitorKeyboard
Seagate Barracuda 320Gb Vista 32bit 32" LCD (Vizio) Logitech G15 
PowerCaseMouse
Corsair HX520w Raidmax Smilodon Logitech MX518 
  hide details  
Reply
Smilodon
(13 items)
 
  
CPUMotherboardGraphicsRAM
e4400 @ 2.9Ghz EVGA 650i Ultra ATX MSI 8800GTS 640 2x1GB G.Skill DDR2 800 
Hard DriveOSMonitorKeyboard
Seagate Barracuda 320Gb Vista 32bit 32" LCD (Vizio) Logitech G15 
PowerCaseMouse
Corsair HX520w Raidmax Smilodon Logitech MX518 
  hide details  
Reply
post #32 of 85
Thread Starter 
Well, I'm in Florida as we speak. I'm having a blast.

Thanks Pauldovi and Mr. Bungle for not flaming me with your opposing arguments, it's much appreciated!

I'll I'm attempting to do is basically tell you, if you can run a higher FSB to match your RAM's frequency, do it. If you can't - then use a divider, because chances are the higher bandwidth on the memory will be beneficial to performance.

I plan on doing more tests, and I look forward to seeing opposing arguments tests. This isn't my laptop I'm using so I won't be able to reply much in the next few days, but thanks again!

Jake
    
CPUMotherboardGraphicsRAM
Intel Q6600 G0 @ 3.6GHz (lapped) EVGA 780i SLI FTW @ 1600MHz FSB 2 x EVGA 8800GTX 768MB (SLI) @ 625/1500/2000MHz 2x2GB G.Skill PC2-8000 PQ @ 800MHz (linked to FSB) 
Hard DriveOptical DriveOSMonitor
2x160GB WD RAID 0, 750GB Samsung Spinpoint F1 LITE-ON 20X SATA DVD/CD RW Windows 7 Ultimate x64 Viewsonic VA2702w 27" LCD 1920x1080 
KeyboardPowerCaseMouse
Logitech Value 100 (beer magnet) Corsair TX 750W Cooler Master ATCS 840 Logitech SBF-90 (cheapest one on Newegg) 
Mouse Pad
Some kinda rubber thing (beer magnet) 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Q6600 G0 @ 3.6GHz (lapped) EVGA 780i SLI FTW @ 1600MHz FSB 2 x EVGA 8800GTX 768MB (SLI) @ 625/1500/2000MHz 2x2GB G.Skill PC2-8000 PQ @ 800MHz (linked to FSB) 
Hard DriveOptical DriveOSMonitor
2x160GB WD RAID 0, 750GB Samsung Spinpoint F1 LITE-ON 20X SATA DVD/CD RW Windows 7 Ultimate x64 Viewsonic VA2702w 27" LCD 1920x1080 
KeyboardPowerCaseMouse
Logitech Value 100 (beer magnet) Corsair TX 750W Cooler Master ATCS 840 Logitech SBF-90 (cheapest one on Newegg) 
Mouse Pad
Some kinda rubber thing (beer magnet) 
  hide details  
Reply
post #33 of 85
Quote:
Originally Posted by Choggs396 View Post
Well, I'm in Florida as we speak. I'm having a blast.

Thanks Pauldovi and Mr. Bungle for not flaming me with your opposing arguments, it's much appreciated!

I'll I'm attempting to do is basically tell you, if you can run a higher FSB to match your RAM's frequency, do it. If you can't - then use a divider, because chances are the higher bandwidth on the memory will be beneficial to performance.

I plan on doing more tests, and I look forward to seeing opposing arguments tests. This isn't my laptop I'm using so I won't be able to reply much in the next few days, but thanks again!

Jake
Hey, choggers is back! What up man? Glad your having a good time down there.

-Oh and sorry for hijacking your thread, but BTW I used your tests to find my max OC... thanks again
Smilodon
(13 items)
 
  
CPUMotherboardGraphicsRAM
e4400 @ 2.9Ghz EVGA 650i Ultra ATX MSI 8800GTS 640 2x1GB G.Skill DDR2 800 
Hard DriveOSMonitorKeyboard
Seagate Barracuda 320Gb Vista 32bit 32" LCD (Vizio) Logitech G15 
PowerCaseMouse
Corsair HX520w Raidmax Smilodon Logitech MX518 
  hide details  
Reply
Smilodon
(13 items)
 
  
CPUMotherboardGraphicsRAM
e4400 @ 2.9Ghz EVGA 650i Ultra ATX MSI 8800GTS 640 2x1GB G.Skill DDR2 800 
Hard DriveOSMonitorKeyboard
Seagate Barracuda 320Gb Vista 32bit 32" LCD (Vizio) Logitech G15 
PowerCaseMouse
Corsair HX520w Raidmax Smilodon Logitech MX518 
  hide details  
Reply
post #34 of 85
Thread Starter 
So, I was browsing around the evga forums I found a couple more tests done by a few guys there that both seem to support my claim (at least somewhat).

The only difference is, they didn't change the multiplier/FSB settings to achieve the same CPU clock - they actually changed the RAM speed/timings by using the divider - that's it. Everything else was the same. And what happened? The system with synced FSB/RAM actually performed better at a lower memory speed. They even mention that the memory bandwidth was of course lower, but for overall performance (which is what 3DMark tests), the slower, synced RAM won. Maybe bandwidth isn't everything.......

Quote:
I ran my DDR2 800 at 5-5-5-15 2t, and I ran again at DDR2 533 at 3-3-3-9 1t, the results were nearly identical. I used 3d Mark ’06. If you run a memory bandwidth test, then the DDR2 800 will win, but I don’t care…I’m after overall performance at the end of the day, and I don’t care how I get there.
and this:
Quote:
In another such comparison I had “bigdawginva” (much thanks) try it with his DDR2 1000 Ram. With an over-clock to 3.37Ghz on an X6800 with DDR 750 at 3-3-3-9 linked/synced he scored 15233 in 3d Mark ’06.

The same system, same over-clock, unlinked DDR 1000 at 4-4-4-12 he scored a 15208. He actually gained a few points trying it linked and synced, but the scores are well within the margin of error. My whole point is that they didn’t drop when running Ram slower than advertised if you adjust your timings accordingly.



He does state that the scores are well within the margin of error. But even seeing that the scores were similar says something, IMO.

Now, it could be just the RAM timings, I admit (and there will probably be some people who will completely blow off the results because of this, sad) - but a better 3Dmark06 score with RAM running 250MHz slower?

Well I liked this test setup so much, I think I'm going to run a similar setup, but with the same RAM timings to see what I get (I still need to do some tests that I was planning on doing weeks ago too, hehe).

The least this says is that memory bandwidth isn't necessarily king when it comes to overall performance.
    
CPUMotherboardGraphicsRAM
Intel Q6600 G0 @ 3.6GHz (lapped) EVGA 780i SLI FTW @ 1600MHz FSB 2 x EVGA 8800GTX 768MB (SLI) @ 625/1500/2000MHz 2x2GB G.Skill PC2-8000 PQ @ 800MHz (linked to FSB) 
Hard DriveOptical DriveOSMonitor
2x160GB WD RAID 0, 750GB Samsung Spinpoint F1 LITE-ON 20X SATA DVD/CD RW Windows 7 Ultimate x64 Viewsonic VA2702w 27" LCD 1920x1080 
KeyboardPowerCaseMouse
Logitech Value 100 (beer magnet) Corsair TX 750W Cooler Master ATCS 840 Logitech SBF-90 (cheapest one on Newegg) 
Mouse Pad
Some kinda rubber thing (beer magnet) 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Q6600 G0 @ 3.6GHz (lapped) EVGA 780i SLI FTW @ 1600MHz FSB 2 x EVGA 8800GTX 768MB (SLI) @ 625/1500/2000MHz 2x2GB G.Skill PC2-8000 PQ @ 800MHz (linked to FSB) 
Hard DriveOptical DriveOSMonitor
2x160GB WD RAID 0, 750GB Samsung Spinpoint F1 LITE-ON 20X SATA DVD/CD RW Windows 7 Ultimate x64 Viewsonic VA2702w 27" LCD 1920x1080 
KeyboardPowerCaseMouse
Logitech Value 100 (beer magnet) Corsair TX 750W Cooler Master ATCS 840 Logitech SBF-90 (cheapest one on Newegg) 
Mouse Pad
Some kinda rubber thing (beer magnet) 
  hide details  
Reply
post #35 of 85
it could also be along the lines of 3dmark isn't testing memory bandwidth that much. i think about what the test is after? GPU performance as a primary thing. there is just one small test for cpu and do we know how much that portion of the test will test the memory on the system?
compy jr
(13 items)
 
  
CPUMotherboardGraphicsHard Drive
xeon 3350 at 3.6ghz Asus P5k-dlx HIS 7970 @ 1150/1900 Lots of them.  
OSMonitorKeyboardPower
vista ultimate x64/xp sp3 compaq 21" junk dell keyboard corsair tx850m 
CaseMouseMouse Pad
upgraded metal box You don't need a mouse Desk 
  hide details  
Reply
compy jr
(13 items)
 
  
CPUMotherboardGraphicsHard Drive
xeon 3350 at 3.6ghz Asus P5k-dlx HIS 7970 @ 1150/1900 Lots of them.  
OSMonitorKeyboardPower
vista ultimate x64/xp sp3 compaq 21" junk dell keyboard corsair tx850m 
CaseMouseMouse Pad
upgraded metal box You don't need a mouse Desk 
  hide details  
Reply
post #36 of 85
Thread Starter 
Quote:
Originally Posted by rx7speed View Post
it could also be along the lines of 3dmark isn't testing memory bandwidth that much. i think about what the test is after? GPU performance as a primary thing. there is just one small test for cpu and do we know how much that portion of the test will test the memory on the system?
3DMark06 is a good determinant of how well your system will perform overall, especially in games. And we can all agree that systems like the ones tested from the EVGA forums I linked to are usually built with gaming in mind.

Maybe you forgot to read this quote:
Quote:
I ran my DDR2 800 at 5-5-5-15 2t, and I ran again at DDR2 533 at 3-3-3-9 1t, the results were nearly identical. I used 3d Mark ’06. If you run a memory bandwidth test, then the DDR2 800 will win, but I don’t care…I’m after overall performance at the end of the day, and I don’t care how I get there.
I'll humor you though and say 3Dmark06 isn't memory bandwidth intensive... so memory speed/timings aside, what's the only other aspect about the two setups that were different then? The divider.

So thanks for pointing that out for me.
    
CPUMotherboardGraphicsRAM
Intel Q6600 G0 @ 3.6GHz (lapped) EVGA 780i SLI FTW @ 1600MHz FSB 2 x EVGA 8800GTX 768MB (SLI) @ 625/1500/2000MHz 2x2GB G.Skill PC2-8000 PQ @ 800MHz (linked to FSB) 
Hard DriveOptical DriveOSMonitor
2x160GB WD RAID 0, 750GB Samsung Spinpoint F1 LITE-ON 20X SATA DVD/CD RW Windows 7 Ultimate x64 Viewsonic VA2702w 27" LCD 1920x1080 
KeyboardPowerCaseMouse
Logitech Value 100 (beer magnet) Corsair TX 750W Cooler Master ATCS 840 Logitech SBF-90 (cheapest one on Newegg) 
Mouse Pad
Some kinda rubber thing (beer magnet) 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Q6600 G0 @ 3.6GHz (lapped) EVGA 780i SLI FTW @ 1600MHz FSB 2 x EVGA 8800GTX 768MB (SLI) @ 625/1500/2000MHz 2x2GB G.Skill PC2-8000 PQ @ 800MHz (linked to FSB) 
Hard DriveOptical DriveOSMonitor
2x160GB WD RAID 0, 750GB Samsung Spinpoint F1 LITE-ON 20X SATA DVD/CD RW Windows 7 Ultimate x64 Viewsonic VA2702w 27" LCD 1920x1080 
KeyboardPowerCaseMouse
Logitech Value 100 (beer magnet) Corsair TX 750W Cooler Master ATCS 840 Logitech SBF-90 (cheapest one on Newegg) 
Mouse Pad
Some kinda rubber thing (beer magnet) 
  hide details  
Reply
post #37 of 85
I read the quote you provided this time and the last. but for 3d itself I really wonder how much ram bandwidth would make a difference in 3dmark since it is more for GPU testing. With this I have noticed that 3dmark is more sensative to GPU changes then it is to cpu speed or memory changes. One thing I have tried once before (can't now as my ram is on rma and so stuck with centon 667 which doesn't oc that well) was run 3dmark at 450*8 and 400*9. the scores are almost the same. I have also tried it at 400*9 and 401*9 same thing almost the same. all of these are done with 1:1 ratio.
this might not sound like much but you have to understand the motherboard I have. it makes a strap change going from 400mhz to 401mhz. this causes a HUGE drop in memory performance when switching straps. this doesn't change the fsb:ram divider or anything like that but rather gives increased latency through the northbridge.
if anything this would either show one of two things. memory performance is just not important at all, or 3dmark just doesn't show any real performance differences due to memory as it doesn't use much for system memory.


if you really want to prove though that the divider hinders performance try the test I gave you before. 450mhz*9 with a 5/4 divider vs 562*6. sure cpu clock speeds are going to be a little different but honestly it won't matter much as unless you are using a p35 board. but at that point there wouldn't be much of a reason to use a divider then.
compy jr
(13 items)
 
  
CPUMotherboardGraphicsHard Drive
xeon 3350 at 3.6ghz Asus P5k-dlx HIS 7970 @ 1150/1900 Lots of them.  
OSMonitorKeyboardPower
vista ultimate x64/xp sp3 compaq 21" junk dell keyboard corsair tx850m 
CaseMouseMouse Pad
upgraded metal box You don't need a mouse Desk 
  hide details  
Reply
compy jr
(13 items)
 
  
CPUMotherboardGraphicsHard Drive
xeon 3350 at 3.6ghz Asus P5k-dlx HIS 7970 @ 1150/1900 Lots of them.  
OSMonitorKeyboardPower
vista ultimate x64/xp sp3 compaq 21" junk dell keyboard corsair tx850m 
CaseMouseMouse Pad
upgraded metal box You don't need a mouse Desk 
  hide details  
Reply
post #38 of 85
Thread Starter 
In all honesty, I really don't have to run the test you provide to prove anything. I understand your motherboard has straps, but the point of these tests aren't to pick out your motherboard or RAM or anything else specifically, or anybodies for that matter... I have to use what I have to do the tests and they've shown quite a lot. Saying anything less, and refusing to take them for what they are, is just refusing to accept the truth simply because you don't like it.

I think you believe I'm trying to tell everyone including you to run your computer in a very specific way. I'm not. I'm just trying to help show people a few examples of why and how different things effect a system using real data and real results. I've said before that some people need to run dividers based on their configurations. I've even said higher memory bandwidth is almost always good. This isn't a guideline or set of rules on how people should run their computer and please don't think it's such.

I understand it's good to take things with a grain of salt, too, but it seems you're purposely trying to make an excuse for everything I present as being wrong, misguided, or irrelevant. You're just hurting yourself by doing this. Look at the data - it's real - and use it to your advantage. Whether or not you apply any of it is up to you, but at least consider it's relevance.
    
CPUMotherboardGraphicsRAM
Intel Q6600 G0 @ 3.6GHz (lapped) EVGA 780i SLI FTW @ 1600MHz FSB 2 x EVGA 8800GTX 768MB (SLI) @ 625/1500/2000MHz 2x2GB G.Skill PC2-8000 PQ @ 800MHz (linked to FSB) 
Hard DriveOptical DriveOSMonitor
2x160GB WD RAID 0, 750GB Samsung Spinpoint F1 LITE-ON 20X SATA DVD/CD RW Windows 7 Ultimate x64 Viewsonic VA2702w 27" LCD 1920x1080 
KeyboardPowerCaseMouse
Logitech Value 100 (beer magnet) Corsair TX 750W Cooler Master ATCS 840 Logitech SBF-90 (cheapest one on Newegg) 
Mouse Pad
Some kinda rubber thing (beer magnet) 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Q6600 G0 @ 3.6GHz (lapped) EVGA 780i SLI FTW @ 1600MHz FSB 2 x EVGA 8800GTX 768MB (SLI) @ 625/1500/2000MHz 2x2GB G.Skill PC2-8000 PQ @ 800MHz (linked to FSB) 
Hard DriveOptical DriveOSMonitor
2x160GB WD RAID 0, 750GB Samsung Spinpoint F1 LITE-ON 20X SATA DVD/CD RW Windows 7 Ultimate x64 Viewsonic VA2702w 27" LCD 1920x1080 
KeyboardPowerCaseMouse
Logitech Value 100 (beer magnet) Corsair TX 750W Cooler Master ATCS 840 Logitech SBF-90 (cheapest one on Newegg) 
Mouse Pad
Some kinda rubber thing (beer magnet) 
  hide details  
Reply
post #39 of 85
I'm not trying to just sit here and say your wrong just to make you wrong.

first test you ran as others have pointed out as well had some flaws in them. for one the FSB was very different which could easily cause problems with lower performance. now what you seem to be doing is using a program that as it seems does not use system memory that much and so yes you are not going to show much for results there.

if you really want to prove something you don't use single test and multiple variables.
try running something that makes a little more use of the system as a whole or use multiple benchmarks. not use a program designed as a gpu test and say "well see system memory speeds don't change teh score so bandwidth might not be so important after all." that was part of the point with my bringing up the straps on my motherboard. it wasn't to nit pick about my board specificly but rather to show that changing that on 3dmark memory itself isn't very memory dependent and as such isn't the best benchmark to use to test memory differences. I know running other benchmarks going from these settings can show a decent difference in performance based upon how memory dependent the program is.

the point I was bringing up also with the two different fsb's one at 450 and one at 562 is that it shows one of the main reasons as to why you would use a divider. which is to optimize the system and max out everything you can. other then tht you really don't want to underclock your fsb just so you can use a divider.


so in the end it's not really what you say I thank you for the attempt. my problem is more on how it is presented and some of the reasons you give for your arguements.
when I get home I will try to show you a little better as to what I'm refering to on this one. again though I am VERY ram limited as max ram speed I can achieve is around 700-740 depending on how nice this damn centon 512mb will be. so my so expect some testing done at substock fsb speeds possibly.
compy jr
(13 items)
 
  
CPUMotherboardGraphicsHard Drive
xeon 3350 at 3.6ghz Asus P5k-dlx HIS 7970 @ 1150/1900 Lots of them.  
OSMonitorKeyboardPower
vista ultimate x64/xp sp3 compaq 21" junk dell keyboard corsair tx850m 
CaseMouseMouse Pad
upgraded metal box You don't need a mouse Desk 
  hide details  
Reply
compy jr
(13 items)
 
  
CPUMotherboardGraphicsHard Drive
xeon 3350 at 3.6ghz Asus P5k-dlx HIS 7970 @ 1150/1900 Lots of them.  
OSMonitorKeyboardPower
vista ultimate x64/xp sp3 compaq 21" junk dell keyboard corsair tx850m 
CaseMouseMouse Pad
upgraded metal box You don't need a mouse Desk 
  hide details  
Reply
post #40 of 85
Thread Starter 
Once again, I revert back to what I said in my previous post. I must, because basically you're saying the same thing you said before. I find it perplexing that you repeat the same criticisms, making me repeat what I've said as well.

If you want to dismiss the tests for the most part, fine. They still present interesting results.... and I never once drew any definite conclusions from them.

Seriously, some of the comments you make are easily construed as "the results mean little because you didn't do this...". Give me a break. Just because I'm not as thorough as you'd like me to be, still doesn't mean the data isn't valid.

You keep telling me to run more tests... while so far you've ran none of your own. Instead of picking apart my data even more, you should try running some.

You borderline flame me about every single aspect of the tests. You could start your own thread with your own results.

I'll try not to be as condescending about your results as you are about mine.
    
CPUMotherboardGraphicsRAM
Intel Q6600 G0 @ 3.6GHz (lapped) EVGA 780i SLI FTW @ 1600MHz FSB 2 x EVGA 8800GTX 768MB (SLI) @ 625/1500/2000MHz 2x2GB G.Skill PC2-8000 PQ @ 800MHz (linked to FSB) 
Hard DriveOptical DriveOSMonitor
2x160GB WD RAID 0, 750GB Samsung Spinpoint F1 LITE-ON 20X SATA DVD/CD RW Windows 7 Ultimate x64 Viewsonic VA2702w 27" LCD 1920x1080 
KeyboardPowerCaseMouse
Logitech Value 100 (beer magnet) Corsair TX 750W Cooler Master ATCS 840 Logitech SBF-90 (cheapest one on Newegg) 
Mouse Pad
Some kinda rubber thing (beer magnet) 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Q6600 G0 @ 3.6GHz (lapped) EVGA 780i SLI FTW @ 1600MHz FSB 2 x EVGA 8800GTX 768MB (SLI) @ 625/1500/2000MHz 2x2GB G.Skill PC2-8000 PQ @ 800MHz (linked to FSB) 
Hard DriveOptical DriveOSMonitor
2x160GB WD RAID 0, 750GB Samsung Spinpoint F1 LITE-ON 20X SATA DVD/CD RW Windows 7 Ultimate x64 Viewsonic VA2702w 27" LCD 1920x1080 
KeyboardPowerCaseMouse
Logitech Value 100 (beer magnet) Corsair TX 750W Cooler Master ATCS 840 Logitech SBF-90 (cheapest one on Newegg) 
Mouse Pad
Some kinda rubber thing (beer magnet) 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
Overclock.net › Forums › Intel › Intel CPUs › FSB and Dividers: My Findings