[Questions] Tightening Primary Memory Timings High Speed DDR3 - Overclock.net - An Overclocking Community

Forum Jump: 

[Questions] Tightening Primary Memory Timings High Speed DDR3

 
Thread Tools
post #1 of 7 (permalink) Old 02-13-2019, 08:39 AM - Thread Starter
New to Overclock.net
 
rsturtle's Avatar
 
Join Date: Feb 2015
Posts: 129
Rep: 11 (Unique: 11)
[Questions] Tightening Primary Memory Timings High Speed DDR3

Hi guys!

I've dabbled with overclocking memory before, but I just picked up a new G. Skill kit, and I want to make the most of my IMC. I have a few questions, but first heres the specs:

16gb (2x8gb) G. Skill 2933 Mhz, 12-14-14-35 @ 1.65v
Msi z97 Gaming 5
4790k @ 4.6/4.1ghz

I have the overclock at the same voltages as normal (should be stable), but to be sure am testing ram with the ring bus and core speed at -100mhz just to be sure. Preliminary testing of the ram seems to show that 2933 @ 12 CL is a no boot, no matter how loose the other timings are. 2800 is also unstable at 12 CL. I can boot 2666 @ 11 CL, so I think that's the best option for this kit.

G. Skill sells a 2666 kit that's 11-13-13-35 @ 1.65v, which also has double sided DIMM's (8gb DIMMS).
This seems like a good stating point. It works. Now I want to tighten the timings.

Questions:
1. When tightening primary timings, does tightening tRCD make it harder to tighten tRP and vice versa? If so which should I prioritize?

2. Is it okay to have tRP be lower than tRCD? I've never seen a kit with timings like that. (I can run 11-13-12, but not 11-12-13)

3. Can memory timings be too loose? For instance is there any reason why I can't run 11-15-15-40 to check if CL 11 at a certain speed is even possible?

4. While it isn't a primary timing, should I prefer T1 command rate and looser timings vs tight timings at T2 command rate, if so to what extent?

These are questions I have been unable to answer by googling and reading guides, and while I could test performance for these myself, I wanted to know if anyone happened to know the answer to any of these questions I have been having. Please feel free to jump into the discussion even if you only have one point to add!

Cheers, thanks!

Processor: i5-4790k @ 4.7Ghz
Motherboard: Z97 MSI Gaming 5
Memory: (RIP) 8gb G.Skill Ripjaws @ 2000Mhz (8-9-9-24 T1) @ 1.55V
Storage: Samsung 970EVO NVME
Power Supply: Corsair CX-600
Graphics: STRIX GTX1080 Bios Flashed...
MSI GTX970 @ 1525/1925
rsturtle is offline  
Sponsored Links
Advertisement
 
post #2 of 7 (permalink) Old 02-13-2019, 02:54 PM
New to Overclock.net
 
Join Date: May 2014
Location: Australia Northern VIC
Posts: 243
Rep: 15 (Unique: 15)
Some of those things can vary from one system to the next grab aida64 give it a few run's at each setting to work out what performs best.
Dasa is offline  
post #3 of 7 (permalink) Old 02-22-2019, 04:31 PM
New to Overclock.net
 
XLifted's Avatar
 
Join Date: Jul 2013
Posts: 1,132
Rep: 36 (Unique: 35)
Quote: Originally Posted by rsturtle View Post
Hi guys!

I've dabbled with overclocking memory before, but I just picked up a new G. Skill kit, and I want to make the most of my IMC. I have a few questions, but first heres the specs:

16gb (2x8gb) G. Skill 2933 Mhz, 12-14-14-35 @ 1.65v
Msi z97 Gaming 5
4790k @ 4.6/4.1ghz

I have the overclock at the same voltages as normal (should be stable), but to be sure am testing ram with the ring bus and core speed at -100mhz just to be sure. Preliminary testing of the ram seems to show that 2933 @ 12 CL is a no boot, no matter how loose the other timings are. 2800 is also unstable at 12 CL. I can boot 2666 @ 11 CL, so I think that's the best option for this kit.

G. Skill sells a 2666 kit that's 11-13-13-35 @ 1.65v, which also has double sided DIMM's (8gb DIMMS).
This seems like a good stating point. It works. Now I want to tighten the timings.

Questions:
1. When tightening primary timings, does tightening tRCD make it harder to tighten tRP and vice versa? If so which should I prioritize?

2. Is it okay to have tRP be lower than tRCD? I've never seen a kit with timings like that. (I can run 11-13-12, but not 11-12-13)

3. Can memory timings be too loose? For instance is there any reason why I can't run 11-15-15-40 to check if CL 11 at a certain speed is even possible?

4. While it isn't a primary timing, should I prefer T1 command rate and looser timings vs tight timings at T2 command rate, if so to what extent?

These are questions I have been unable to answer by googling and reading guides, and while I could test performance for these myself, I wanted to know if anyone happened to know the answer to any of these questions I have been having. Please feel free to jump into the discussion even if you only have one point to add!

Cheers, thanks!

Just because you can boot it, does NOT mean you can run it. Do a MemTest64, and if you start seeing errors in program, or blue screens...you know you pushed an imaginary clock that the ram can NOT maintain.

I have been checking my stuff that way too
XLifted is offline  
Sponsored Links
Advertisement
 
post #4 of 7 (permalink) Old 02-22-2019, 04:56 PM
Iconoclast
 
Blameless's Avatar
 
Join Date: Feb 2008
Posts: 29,972
Rep: 3131 (Unique: 1869)
If at all possible, find out the exact memory ICs you'll be using. If you already have the memory, AIDA64 or anything else that can read detailed SPD info should reveal this. If you don't have the memory look up the exact model number and revision.

Quote: Originally Posted by rsturtle View Post
1. When tightening primary timings, does tightening tRCD make it harder to tighten tRP and vice versa? If so which should I prioritize?
Depends on the ICs used, but these timings are usually fairly independent stability wise. However it's hard to predict which one can be tightened more without experience with similar ICs.

Quote: Originally Posted by rsturtle View Post
2. Is it okay to have tRP be lower than tRCD? I've never seen a kit with timings like that. (I can run 11-13-12, but not 11-12-13)
Tighter tRP is more common than tRCD, but I've definitely had some DDR3 that can run lower tRCD than tRP.

Quote: Originally Posted by rsturtle View Post
3. Can memory timings be too loose? For instance is there any reason why I can't run 11-15-15-40 to check if CL 11 at a certain speed is even possible?
A few memory timings are dependent on others and loosening dependencies can occasionally cause issues, but in general, it is a good idea to loosen everything else up to see if a specific tighter timing is viable.

Quote: Originally Posted by rsturtle View Post
4. While it isn't a primary timing, should I prefer T1 command rate and looser timings vs tight timings at T2 command rate, if so to what extent?
Command rate 1T vs. 2T is typically about the same rough performance impact as increasing CAS by one. If you have to do more than increase the primaries by one each, the tighter CMDR rate is almost certainly not worthwhile.

...rightful liberty is unobstructed action according to our will within limits drawn around us by the equal rights of others. I do not add 'within the limits of the law,' because law is often but the tyrant's will, and always so when it violates the right of an individual. -- Thomas Jefferson
Haswell-E
(15 items)
Broadwell-E
(11 items)
Westmere
(13 items)
CPU
5820K @ 4.25/4.125GHz core/uncore, 1.21/1.2v
Motherboard
Gigabyte X99 SOC Champion (F23c + updated microcode)
GPU
Gigabyte AORUS GTX 1080 Ti (F3P) @ 2025/1485(5940), 1.031v
RAM
4x4GiB Crucial @ 2667, 12-11-12-26-T1, 1.39v
Hard Drive
Plextor M6e 128GB (fw 1.06) M.2 (PCI-E 2.0 2x)
Hard Drive
2x Crucial M4 256GB
Hard Drive
4x WD Scorpio Black 500GB
Power Supply
Seasonic Focus Plus 850 Platinum (SSR-850PX)
Cooling
Noctua NH-D15S + 2x Thermalright TY-143
Case
Corsair Carbide Air 540
Operating System
Windows 7 Professional x64 SP1
Monitor
BenQ BL3200PT
Keyboard
Filco Majestouch Tenkeyless (MX Brown)
Mouse
Logitech G402
Audio
Realtek ALC1150 + M-Audio AV40
CPU
i7-6800K @ 4.2(4.1)/3.4GHz core(AVX)/uncore, 1.345/1.15v
Motherboard
ASRock X99 OC Formula (P3.40)
GPU
MSI GTX 1080 Ti Duke OC @ 1950/1485MHz, 1.043v
RAM
4x4GiB Crucial DDR4-2400 @ 11-13-12-28-T2, 1.33v
Hard Drive
Intel 600p 256GB NVMe
Hard Drive
2x HGST Travelstar 7k1000 1TB
Power Supply
Corsair RM1000x
Cooling
Thermalright True Spirit 140 Direct + TY-143
Case
Fractal Design Define R4
Operating System
Windows Server 2016 Datacenter
Audio
ASUS Xonar DX
CPU
X5670 @ 4.3/3.7GHz core/uncore, 1.36 vcore, 1.35 qpi/vtt
Motherboard
Gigabyte X58A-UD5 r2.0 w/FF3mod10 BIOS
GPU
Sapphire Fury Nitro OC+ @ 1080/545, 1.275vGPU/1.394vHBM
RAM
2x Samsung MV-3V4G3D/US @ 2050, 10-11-11-30-T1, 1.52v
RAM
1x Crucial BLT4G3D1608ET3LX0 @ 2050, 10-11-11-30-T1, 1.52v
Hard Drive
OCZ (Toshiba) Trion 150 120GB
Hard Drive
Hyundai Sapphire 120GB
Hard Drive
3x Hitachi Deskstar 7k1000.C 1TB
Power Supply
Seasonic SS-860XP2
Cooling
Noctua NH-D14
Case
Fractal Design Define R5
Operating System
Windows 7 Pro x64 SP1
Audio
ASUS Xonar DS
▲ hide details ▲
Blameless is offline  
post #5 of 7 (permalink) Old 02-23-2019, 10:14 AM - Thread Starter
New to Overclock.net
 
rsturtle's Avatar
 
Join Date: Feb 2015
Posts: 129
Rep: 11 (Unique: 11)
Quote: Originally Posted by XLifted View Post
Just because you can boot it, does NOT mean you can run it. Do a MemTest64, and if you start seeing errors in program, or blue screens...you know you pushed an imaginary clock that the ram can NOT maintain.

I have been checking my stuff that way too
This is very true. I found out the settings I had listed earlier would run fine for hours and then crash on memtest. I see it crash out around 400-600% coverage. It's frustratingly close. I might give voltages a tiny bump and see if that stabilizes the clock speeds.

Quote: Originally Posted by Blameless View Post
If at all possible, find out the exact memory ICs you'll be using. If you already have the memory, AIDA64 or anything else that can read detailed SPD info should reveal this. If you don't have the memory look up the exact model number and revision.



Depends on the ICs used, but these timings are usually fairly independent stability wise. However it's hard to predict which one can be tightened more without experience with similar ICs.



Tighter tRP is more common than tRCD, but I've definitely had some DDR3 that can run lower tRCD than tRP.



A few memory timings are dependent on others and loosening dependencies can occasionally cause issues, but in general, it is a good idea to loosen everything else up to see if a specific tighter timing is viable.



Command rate 1T vs. 2T is typically about the same rough performance impact as increasing CAS by one. If you have to do more than increase the primaries by one each, the tighter CMDR rate is almost certainly not worthwhile.
Thanks for the responses! Especially point number four, that's what I had suspected but I wasn't sure.

Processor: i5-4790k @ 4.7Ghz
Motherboard: Z97 MSI Gaming 5
Memory: (RIP) 8gb G.Skill Ripjaws @ 2000Mhz (8-9-9-24 T1) @ 1.55V
Storage: Samsung 970EVO NVME
Power Supply: Corsair CX-600
Graphics: STRIX GTX1080 Bios Flashed...
MSI GTX970 @ 1525/1925
rsturtle is offline  
post #6 of 7 (permalink) Old 02-23-2019, 10:16 AM - Thread Starter
New to Overclock.net
 
rsturtle's Avatar
 
Join Date: Feb 2015
Posts: 129
Rep: 11 (Unique: 11)
Adding this for completeness, I got another response to this question when I cross-posted this on reddit. Here it is below: (username varexos717)

"1. No, it shouldn't. The TL;DR is that the operations tRP and tRCD correspond to don't overlap; they're mutually exclusive as far as I understand it. There are some timings which are related in the way you're thinking (e.g. tCL/tCWL, tRDWR/tWRRD), but from my experience and research, it doesn't apply to these two.

2. Yes, see above. Also, I'm pretty sure some Intel platforms have a tRCD/RP timing that's the two combined, so I would guess that's the reason these timings are often set the same...don't quote me on that though.

3. I don't know the technicalities of this one, but yes. There's a reason there aren't kits rated for, I don't know, 4000CL30 or something crazy like that. There's kind of a sweet spot for your timings, and just like setting them too low causes instability, setting them too high can also cause instability.

4. The difference between 2T and 1T is basically that at 2T, your memory essentially has an extra chance to receive a signal from your CPU, which is why that helps improve stability vs 1T. That might lead you to think that your memory runs twice as slow, but really the difference between the two is very small. Off the top of my head, AMD's internal data for Ryzen, for example, shows that you might gain a couple of FPS in high refresh rate gaming with 1T, and remember Ryzen is very memory sensitive. So TL;DR if 2T lets you run your memory at higher frequency/tighter timings, there's no real reason not to use it. But if it doesn't help, then 1T is ideal of course."

Processor: i5-4790k @ 4.7Ghz
Motherboard: Z97 MSI Gaming 5
Memory: (RIP) 8gb G.Skill Ripjaws @ 2000Mhz (8-9-9-24 T1) @ 1.55V
Storage: Samsung 970EVO NVME
Power Supply: Corsair CX-600
Graphics: STRIX GTX1080 Bios Flashed...
MSI GTX970 @ 1525/1925
rsturtle is offline  
post #7 of 7 (permalink) Old 02-26-2019, 10:41 AM
Hardware Rep
 
GSKILL SUPPORT's Avatar
 
Join Date: Nov 2016
Posts: 52
Rep: 5 (Unique: 5)
Some times too low of timings may not perform as well, so it's always necessary to test and see what combination works best. As Blameless mentioned, IC chip characteristics vary with brand, model, batch, etc.

GSKILL SUPPORT is offline  
Reply

Tags
Ddr3 , Memory , primary timings , ram overclocking , tighten timings

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off