Jump to content

Modify memory manually vs setting docp and modifying it

Go to solution Solved by nick name,
1 minute ago, lotus10101 said:

Interesting, so this droop in bclk while in a docp profile could explain why I can use 1T but when I'm not in docp and set the exact same timings same voltage I cannot and it will crash at startup unless I put voltage I'm not real comfortable with lol

Hmmmm.  A small drop in BCLK wouldn't explain that, but maybe Spread Spectrum would?  

 

What mobo do you have?

If I set a docp, then later modified my timings manually but still had a docp profile set will I get different results if I were to have no docp but have the same timings?

                          Ryzen 5800X3D(Because who doesn't like a phat stack of cache?) GPU - 7700Xt

                                                           X470 Strix f gaming, 32GB Corsair vengeance, WD Blue 500GB NVME-WD Blue2TB HDD, 700watts EVGA Br

 ~Extra L3 cache is exciting, every time you load up a new game or program you never know what your going to get, will it perform like a 5700x or are we beating the 14900k today? 😅~

Link to post
Share on other sites

Benchmark or real world performance, basically I had originally set a Docp and then later started to modify my ram timings manually but didn't realize I still had a docp profile set and I'm wondering if it would be worth it to reset my bios to default and re input my current timings or would overall performance be the same as it is now

                          Ryzen 5800X3D(Because who doesn't like a phat stack of cache?) GPU - 7700Xt

                                                           X470 Strix f gaming, 32GB Corsair vengeance, WD Blue 500GB NVME-WD Blue2TB HDD, 700watts EVGA Br

 ~Extra L3 cache is exciting, every time you load up a new game or program you never know what your going to get, will it perform like a 5700x or are we beating the 14900k today? 😅~

Link to post
Share on other sites

2 minutes ago, Benji said:

Considering the fact that the integrated XMP profiles (or, as they're also called on AMD platforms, DOCP) have the exact values that the manufacturer also puts out on their websites, I think that manually setting them will make literally no difference compared to using the pre-programmed profile. I heard multiple times that it supposedly performs better, but I think it's just a telltale.

I definitely got better performance manually tightening the timings but I still had the docp profile set when I started changing timings,I'm just trying to figure out if there would be some difference if I deactivated the docp before changing timings

                          Ryzen 5800X3D(Because who doesn't like a phat stack of cache?) GPU - 7700Xt

                                                           X470 Strix f gaming, 32GB Corsair vengeance, WD Blue 500GB NVME-WD Blue2TB HDD, 700watts EVGA Br

 ~Extra L3 cache is exciting, every time you load up a new game or program you never know what your going to get, will it perform like a 5700x or are we beating the 14900k today? 😅~

Link to post
Share on other sites

IMG_20200815_120807.thumb.jpg.a83f1f77bb76c71ed1b1c38fb7c2a585.jpgThis is the profile currently set in my biosIMG_20200815_120538.thumb.jpg.567676011e26a4194a2be2f228da614f.jpgIMG_20200815_120538.thumb.jpg.567676011e26a4194a2be2f228da614f.jpgthese are my current ram timings

                          Ryzen 5800X3D(Because who doesn't like a phat stack of cache?) GPU - 7700Xt

                                                           X470 Strix f gaming, 32GB Corsair vengeance, WD Blue 500GB NVME-WD Blue2TB HDD, 700watts EVGA Br

 ~Extra L3 cache is exciting, every time you load up a new game or program you never know what your going to get, will it perform like a 5700x or are we beating the 14900k today? 😅~

Link to post
Share on other sites

The Cas latency is the main timing that would bring better performance.

Trfc will bring some performance..

Trc will bring better performance.

 

All together, you may decrease latency by a few ns, but you're not really increasing bandwidth unless you overclock the cpu or overclock the ram or both together.

 

Some of the best testing can be done through AIDA64 memory and cache benchmark, but real world performance won't really be that noticeable

 

But as long as you measure the tweaks with something, that's good. 

performance gains are performance gains I always say, no matter how small. 

Link to post
Share on other sites

The only reason I haven't OC my CPU much is because I have 212 Evo 🤣 so it reaches 72° at 4.15hz  in cinebench r20, and I heard that ryzen fclock will decouple after 3600mhz So I haven't tried to go beyond

                          Ryzen 5800X3D(Because who doesn't like a phat stack of cache?) GPU - 7700Xt

                                                           X470 Strix f gaming, 32GB Corsair vengeance, WD Blue 500GB NVME-WD Blue2TB HDD, 700watts EVGA Br

 ~Extra L3 cache is exciting, every time you load up a new game or program you never know what your going to get, will it perform like a 5700x or are we beating the 14900k today? 😅~

Link to post
Share on other sites

2 minutes ago, lotus10101 said:

The only reason I haven't OC my CPU much is because I have 212 Evo 🤣 so it reaches 72° at 4.15hz  in cinebench r20, and I heard that ryzen fclock will decouple after 3600mhz So I haven't tried to go beyond

Most chips 3800mhz is about the fastest it'll go. 

But what does that mean? 

You unlinked the IF and push the memory faster but will get diminishing returns after 3800mhz frequency. That happens because you no longer gain performance from the IF because it's not getting overclocked beyond that frequency. 

Link to post
Share on other sites

1 minute ago, ShrimpBrime said:

Most chips 3800mhz is about the fastest it'll go. 

But what does that mean? 

You unlinked the IF and push the memory faster but will get diminishing returns after 3800mhz frequency. That happens because you no longer gain performance from the IF because it's not getting overclocked beyond that frequency. 

So I should be able to push my memory frequency higher,and still have 1 to 1 with my IF? 

                          Ryzen 5800X3D(Because who doesn't like a phat stack of cache?) GPU - 7700Xt

                                                           X470 Strix f gaming, 32GB Corsair vengeance, WD Blue 500GB NVME-WD Blue2TB HDD, 700watts EVGA Br

 ~Extra L3 cache is exciting, every time you load up a new game or program you never know what your going to get, will it perform like a 5700x or are we beating the 14900k today? 😅~

Link to post
Share on other sites

4 minutes ago, ShrimpBrime said:

Most chips 3800mhz is about the fastest it'll go. 

But what does that mean? 

You unlinked the IF and push the memory faster but will get diminishing returns after 3800mhz frequency. That happens because you no longer gain performance from the IF because it's not getting overclocked beyond that frequency. 

I'm not sure if I should try it though,I'm already 1.41v on dram to get CL15 ,not sure more would be safe for day to day use?

                          Ryzen 5800X3D(Because who doesn't like a phat stack of cache?) GPU - 7700Xt

                                                           X470 Strix f gaming, 32GB Corsair vengeance, WD Blue 500GB NVME-WD Blue2TB HDD, 700watts EVGA Br

 ~Extra L3 cache is exciting, every time you load up a new game or program you never know what your going to get, will it perform like a 5700x or are we beating the 14900k today? 😅~

Link to post
Share on other sites

3 minutes ago, lotus10101 said:

So I should be able to push my memory frequency higher,and still have 1 to 1 with my IF? 

Overclocking isnt a promise. 

If your chip can run a higher IF frequency, yes.

Can still unlinked the IF and just push the memory also. 

But yes.

Link to post
Share on other sites

Just now, lotus10101 said:

I'm not sure if I should try it though,I'm already 1.41v on dram to get CL15 ,not sure more would be safe for day to day use?

That's what testing is all about.

So you loosen cas to 16. Maintain the 1.410v.

That's fine. You won't loose much latency, you'll gain it back from the overclock.

 

I dont usually run my B-Die kits past 1.60v actively cooled.

Monitoring RAM temps is a thing you've know...

Link to post
Share on other sites

How do you monitor the ram temp? They don't have sensors, did you install a sensor?

                          Ryzen 5800X3D(Because who doesn't like a phat stack of cache?) GPU - 7700Xt

                                                           X470 Strix f gaming, 32GB Corsair vengeance, WD Blue 500GB NVME-WD Blue2TB HDD, 700watts EVGA Br

 ~Extra L3 cache is exciting, every time you load up a new game or program you never know what your going to get, will it perform like a 5700x or are we beating the 14900k today? 😅~

Link to post
Share on other sites

When you set DOCP it changes values once.  You can change values afterwards.  If you go back in BIOS and switch from DOCP and back to DOCP it will change the values.  Leaving it on won't alter values you've entered manually.

 

I don't know what board you have, but if it's like mine then setting DOCP will cause BCLK to droop to 99.8 from 100 (might be a spread spectrum thing).  So in that sense -- using DOCP reduces performance.  As such -- I set to Manual and enter all values manually.  

 

And that 3600C15 kit of RAM you have will almost certainly do 3800MHz at 14-15-14-14 at 1T with GDM enabled.  Depending on how tight your other timings are will determine the voltage required.  Just make sure to also set the appropriate FCLK in BIOS.  Actually, always set it in BIOS and never leave it to Auto as it might not get set which I think is what we're seeing in your screenshots.  

AMD Ryzen 5800XFractal Design S36 360 AIO w/6 Corsair SP120L fans  |  Asus Crosshair VII WiFi X470  |  G.SKILL TridentZ 4400CL19 2x8GB @ 3800MHz 14-14-14-14-30  |  EVGA 3080 FTW3 Hybrid  |  Samsung 970 EVO M.2 NVMe 500GB - Boot Drive  |  Samsung 850 EVO SSD 1TB - Game Drive  |  Seagate 1TB HDD - Media Drive  |  EVGA 650 G3 PSU | Thermaltake Core P3 Case 

Link to post
Share on other sites

4 hours ago, ShrimpBrime said:

Better memory kits have sensors. Yours probably do not.

I use HWInfo64 to monitor temps.

It looks like he has a 3600C15 kit and I don't recall seeing anyone but G.Skill selling them.  If it's a TridentZ kit then he has sensors. 

 

I really don't get why higher-end kits don't and are marketed to overclockers (looking at you 8-pack b-die kit).  The Viper kits I can forgive because they're so cheap, but you can get equivalent performance from the 3600C15 TridentZ kit for less money and with a sensor.  

AMD Ryzen 5800XFractal Design S36 360 AIO w/6 Corsair SP120L fans  |  Asus Crosshair VII WiFi X470  |  G.SKILL TridentZ 4400CL19 2x8GB @ 3800MHz 14-14-14-14-30  |  EVGA 3080 FTW3 Hybrid  |  Samsung 970 EVO M.2 NVMe 500GB - Boot Drive  |  Samsung 850 EVO SSD 1TB - Game Drive  |  Seagate 1TB HDD - Media Drive  |  EVGA 650 G3 PSU | Thermaltake Core P3 Case 

Link to post
Share on other sites

1 minute ago, nick name said:

It looks like he has a 3600C15 kit and I don't recall seeing anyone but G.Skill selling them.  If it's a TridentZ kit then he has sensors. 

 

I really don't get why higher-end kits don't and are marketed to overclockers (looking at you 8-pack b-die kit).  The Viper kits I can forgive because they're so cheap, but you can get equivalent performance from the 3600C15 TridentZ kit for less money and with a sensor.  

Right, for sure.

I dont go high voltage often. 

Even benching 4000 CL 16 is fine unless the benchmark is strictly cpu and memory bound.

 

So IMO CL 17 at 4000 is fine too. There's plenty of bandwidth for gaming, the timings become less needed to be tight.

Link to post
Share on other sites

Just now, ShrimpBrime said:

Right, for sure.

I dont go high voltage often. 

Even benching 4000 CL 16 is fine unless the benchmark is strictly cpu and memory bound.

 

So IMO CL 17 at 4000 is fine too. There's plenty of bandwidth for gaming, the timings become less needed to be tight.

I wish I had another Intel system I could use for RAM overclocking.  I can overclock with my 3900X, but it's just not the same after you break the 1:1.  I've also learned that it takes stress off the IMC when you break 1:1 so it doesn't give me a true representation of what the RAM kit is actually capable of.  If that makes any sense to anyone but me.  

AMD Ryzen 5800XFractal Design S36 360 AIO w/6 Corsair SP120L fans  |  Asus Crosshair VII WiFi X470  |  G.SKILL TridentZ 4400CL19 2x8GB @ 3800MHz 14-14-14-14-30  |  EVGA 3080 FTW3 Hybrid  |  Samsung 970 EVO M.2 NVMe 500GB - Boot Drive  |  Samsung 850 EVO SSD 1TB - Game Drive  |  Seagate 1TB HDD - Media Drive  |  EVGA 650 G3 PSU | Thermaltake Core P3 Case 

Link to post
Share on other sites

1 minute ago, nick name said:

I wish I had another Intel system I could use for RAM overclocking.  I can overclock with my 3900X, but it's just not the same after you break the 1:1.  I've also learned that it takes stress off the IMC when you break 1:1 so it doesn't give me a true representation of what the RAM kit is actually capable of.  If that makes any sense to anyone but me.  

Think about it. 

Once you unlinked the IF the memory gains are more similar to Intel's.

 

 

Link to post
Share on other sites

Just now, ShrimpBrime said:

Think about it. 

Once you unlinked the IF the memory gains are more similar to Intel's.

 

 

Well I make mention because I found that my newer b-die kits can't do 1T GDM off like my older kit can.  So when I played with 3800C15 at 1T GDM off and seeing it be unstable confused me because at 4200C15 1T GDM off it was perfectly fine.  Then I realized that at 4200 it wasn't running 1:1 so I re-tested 3800C15 1T GDM off with the IF set at 1800MHz instead of 1900MHz (1:1) and that ran no problem.  

 

So I'm not sure if on an Intel rig the kit would run 4200C15 1T or if it would need 2T.  

 

I bought a 4400C19 kit because I'm assuming the new Ryzen 4000 CPUs will run at least IF 2000MHz and maybe 2100MHz.  So I've been playing with it a lot lately.  

AMD Ryzen 5800XFractal Design S36 360 AIO w/6 Corsair SP120L fans  |  Asus Crosshair VII WiFi X470  |  G.SKILL TridentZ 4400CL19 2x8GB @ 3800MHz 14-14-14-14-30  |  EVGA 3080 FTW3 Hybrid  |  Samsung 970 EVO M.2 NVMe 500GB - Boot Drive  |  Samsung 850 EVO SSD 1TB - Game Drive  |  Seagate 1TB HDD - Media Drive  |  EVGA 650 G3 PSU | Thermaltake Core P3 Case 

Link to post
Share on other sites

1 hour ago, nick name said:

Well I make mention because I found that my newer b-die kits can't do 1T GDM off like my older kit can.  So when I played with 3800C15 at 1T GDM off and seeing it be unstable confused me because at 4200C15 1T GDM off it was perfectly fine.  Then I realized that at 4200 it wasn't running 1:1 so I re-tested 3800C15 1T GDM off with the IF set at 1800MHz instead of 1900MHz (1:1) and that ran no problem.  

 

So I'm not sure if on an Intel rig the kit would run 4200C15 1T or if it would need 2T.  

 

I bought a 4400C19 kit because I'm assuming the new Ryzen 4000 CPUs will run at least IF 2000MHz and maybe 2100MHz.  So I've been playing with it a lot lately.  

Could be 2 different PCBs you have. It's not just about the ICs being used. 

 

You can find the difference with this nifty picture.

 

A1 on the left and A2 on the right.

 

 

 

1418112903_a2vsa1.thumb.jpg.02ebcc4fa845cc761b1de6786c7035d5.jpg

Link to post
Share on other sites

3 hours ago, nick name said:

When you set DOCP it changes values once.  You can change values afterwards.  If you go back in BIOS and switch from DOCP and back to DOCP it will change the values.  Leaving it on won't alter values you've entered manually.

 

I don't know what board you have, but if it's like mine then setting DOCP will cause BCLK to droop to 99.8 from 100 (might be a spread spectrum thing).  So in that sense -- using DOCP reduces performance.  As such -- I set to Manual and enter all values manually.  

 

And that 3600C15 kit of RAM you have will almost certainly do 3800MHz at 14-15-14-14 at 1T with GDM enabled.  Depending on how tight your other timings are will determine the voltage required.  Just make sure to also set the appropriate FCLK in BIOS.  Actually, always set it in BIOS and never leave it to Auto as it might not get set which I think is what we're seeing in your screenshots.  

According to the package it's 3600mhz CL17 it's cheap Adata XPG Bdie stuff,I just bought it because I heard it would do better than my hijinx 3200 MHz @CL14 ,this Bdie XPG kit will do CL14  3600mhz but only at 1.45v but crash in games and windows crashes as soon as it loads using 1T,atm I can only get stable with no apperant errors @ 3600mhz CL15 2T, and I'd rather not use 1.5v to force it because I can't find two people on the net who can agree if 1.5v is safe or not for daily use

                          Ryzen 5800X3D(Because who doesn't like a phat stack of cache?) GPU - 7700Xt

                                                           X470 Strix f gaming, 32GB Corsair vengeance, WD Blue 500GB NVME-WD Blue2TB HDD, 700watts EVGA Br

 ~Extra L3 cache is exciting, every time you load up a new game or program you never know what your going to get, will it perform like a 5700x or are we beating the 14900k today? 😅~

Link to post
Share on other sites

3 minutes ago, lotus10101 said:

According to the package it's 3600mhz CL17 it's cheap Adata XPG Bdie stuff,I just bought it because I heard it would do better than my hijinx 3200 MHz @CL14 ,this Bdie XPG kit will do CL14  3600mhz but only at 1.45v but crash in games and windows crashes as soon as it loads using 1T,atm I can only get stable with no apperant errors @ 3600mhz CL15 2T, and I'd rather not use 1.5v to force it higher because I can't find two people on the net who can agree if 1.5v is safe or not for daily use

 

                          Ryzen 5800X3D(Because who doesn't like a phat stack of cache?) GPU - 7700Xt

                                                           X470 Strix f gaming, 32GB Corsair vengeance, WD Blue 500GB NVME-WD Blue2TB HDD, 700watts EVGA Br

 ~Extra L3 cache is exciting, every time you load up a new game or program you never know what your going to get, will it perform like a 5700x or are we beating the 14900k today? 😅~

Link to post
Share on other sites

3 hours ago, nick name said:

When you set DOCP it changes values once.  You can change values afterwards.  If you go back in BIOS and switch from DOCP and back to DOCP it will change the values.  Leaving it on won't alter values you've entered manually.

 

I don't know what board you have, but if it's like mine then setting DOCP will cause BCLK to droop to 99.8 from 100 (might be a spread spectrum thing).  So in that sense -- using DOCP reduces performance.  As such -- I set to Manual and enter all values manually.  

 

And that 3600C15 kit of RAM you have will almost certainly do 3800MHz at 14-15-14-14 at 1T with GDM enabled.  Depending on how tight your other timings are will determine the voltage required.  Just make sure to also set the appropriate FCLK in BIOS.  Actually, always set it in BIOS and never leave it to Auto as it might not get set which I think is what we're seeing in your screenshots.  

Interesting, so this droop in bclk while in a docp profile could explain why I can use 1T but when I'm not in docp and set the exact same timings same voltage I cannot and it will crash at startup unless I put voltage I'm not real comfortable with lol

                          Ryzen 5800X3D(Because who doesn't like a phat stack of cache?) GPU - 7700Xt

                                                           X470 Strix f gaming, 32GB Corsair vengeance, WD Blue 500GB NVME-WD Blue2TB HDD, 700watts EVGA Br

 ~Extra L3 cache is exciting, every time you load up a new game or program you never know what your going to get, will it perform like a 5700x or are we beating the 14900k today? 😅~

Link to post
Share on other sites

1 minute ago, lotus10101 said:

Interesting, so this droop in bclk while in a docp profile could explain why I can use 1T but when I'm not in docp and set the exact same timings same voltage I cannot and it will crash at startup unless I put voltage I'm not real comfortable with lol

Hmmmm.  A small drop in BCLK wouldn't explain that, but maybe Spread Spectrum would?  

 

What mobo do you have?

AMD Ryzen 5800XFractal Design S36 360 AIO w/6 Corsair SP120L fans  |  Asus Crosshair VII WiFi X470  |  G.SKILL TridentZ 4400CL19 2x8GB @ 3800MHz 14-14-14-14-30  |  EVGA 3080 FTW3 Hybrid  |  Samsung 970 EVO M.2 NVMe 500GB - Boot Drive  |  Samsung 850 EVO SSD 1TB - Game Drive  |  Seagate 1TB HDD - Media Drive  |  EVGA 650 G3 PSU | Thermaltake Core P3 Case 

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×