Jump to content

14" Dell Latitude E6430 - Performance Upgrades and System Mods


Recommended Posts

Oh nice. Does this also mean that you can theoretically have eGPU SLI: x2.2 with the mPCIe slots, and x2.2 with the ODD/Sata+EC?

As Tech Inferno Fan said SLI is not possible.

On the other hand crossfire would be possible with older cards and 13.4 legacy catalyst in windows as shown here http://forum.techinferno.com/diy-e-gpu-projects/5622-%5Bblog%5D-crossfire-testing-3.html#post83563

Link to comment
Share on other sites

Thank you guys. So x4.2 is the closest we can hope for then.

How do you enable the ASPMs via UEFI variables? Planning also to migrate to Windows 10 so I might go the UEFI route too over MBR+Setup 1.30.

Setting: ASPM Support, Variable: 0xC 
Option: Disabled, Value: 0x0
Option: Auto, Value: 0x37
Option: Force L0s, Value: 0x1

I think I set this to Auto from Disabled.

Native ASPM, Variable: 0xB04
Option: Enabled, Value: 0x1
Option: Disabled, Value: 0x0

and this to Enabled from Disabled.

I will check on the next reboot.

  • Thumbs Up 1
Link to comment
Share on other sites

Thank you @timohour,

I read the [GUIDE] Dell E6530 CPU TDP/multi unlocking and am trying to understand the actual steps involved.

My (Simplified) Steps in CPU TDP/Multi Unlocking:

  1. Obtain the IFR file (E6430-A07_IFR.txt)
  2. Obtain the NVRAM tool (http://brains.by/posts/bootx64.7z)
  3. Format a USB thumb drive to FAT32 and place the bootx64.efi file in the EFI\Boot\ folder.
  4. Boot to EFI on Startup (Pressing F12)
  5. Load the bootx64.efi (Loads Grub)
  6. Overclock by assigning new values to certain NVRAM addresses (using the "setup_var address value" command)
  7. Reboot

Just a few questions regarding step 6 that I hope you don't mind answering..

  • How would I know what the ideal core limit is? Do I assign in increments and do trial and error, or should I be watching out for something here?
  • What is that adjusting/balancing of multiplier in relation to the BCLK OC headroom all about?
  • Can I use the 99% CPU usage setting to prevent the overclock when not needing max performance?

So sorry, but I just couldn't follow how you chose 1C=40, 2C=39, 4C=38 as your stable core limits, moreso that part with BCLK OC headroom in your post.

Thanks a lot!

Link to comment
Share on other sites

With the dynamic TOLUD I could see that TOLUD would go as low as 3.25GB. (at least on my Latitude). This wasn't enough to use dGPU and eGPU at the same time without compaction or connect multiple eGPUs. That was at least in my case.

Mine has gone as low as 2.75GB with iGPU, dGPU, and eGPU on and allocated. Nando4 had problems getting his to lower though and it didn't do it until he changed his memory config. I think what you need to do is set to do a thorough boot every boot in BIOS. In nando's case it did not do a thorough boot until it detected a memory change.

Link to comment
Share on other sites

Thank you @timohour,

I read the [GUIDE] Dell E6530 CPU TDP/multi unlocking and am trying to understand the actual steps involved.

My (Simplified) Steps in CPU TDP/Multi Unlocking:

  1. Obtain the IFR file (E6430-A07_IFR.txt)
  2. Obtain the NVRAM tool (http://brains.by/posts/bootx64.7z)
  3. Format a USB thumb drive to FAT32 and place the bootx64.efi file in the EFI\Boot\ folder.
  4. Boot to EFI on Startup (Pressing F12)
  5. Load the bootx64.efi (Loads Grub)
  6. Overclock by assigning new values to certain NVRAM addresses (using the "setup_var address value" command)
  7. Reboot

Just a few questions regarding step 6 that I hope you don't mind answering..

  • How would I know what the ideal core limit is? Do I assign in increments and do trial and error, or should I be watching out for something here?
  • What is that adjusting/balancing of multiplier in relation to the BCLK OC headroom all about?
  • Can I use the 99% CPU usage setting to prevent the overclock when not needing max performance?

So sorry, but I just couldn't follow how you chose 1C=40, 2C=39, 4C=38 as your stable core limits, moreso that part with BCLK OC headroom in your post.

Thanks a lot!

Answering your questions:

* You wouldn't know what the ideal core limit is. On the i7-37x0/38x0QMs though you can set only 4 more bins for each turbo. Even if you set a bigger number you can't surpass your chip's maximum.

e.g. an OEM i7-3720QM would have the following original Turbo Boosts:

1-core --> 3.6GHz --> (x36)

2-core --> 3.5GHz --> (x35)

3-core --> 3.4GHz --> (x34)

4-core --> 3.4GHz --> (x34)

Using the maximum 4 more multipliers Intel unlocked on 37x0/38x0 on the 3720QM you can get

1-core --> 4.0GHz --> (x40)

2-core --> 3.9GHz --> (x39)

3-core --> 3.8GHz --> (x38)

4-core --> 3.8GHz --> (x38)

That's how I chose the multipliers. On your situation with the i7-3840QM your maximum multiplier would be

1-core --> 4.2GHz --> (x42)

2-core --> 4.1GHz --> (x41)

3-core --> 4.0GHz --> (x40)

4-core --> 4.0GHz --> (x40)

I can't tell if it would boot with these settings though so I would suggest that you use BCLK OC on lower multipliers to make sure you can use that high clocks because if you don't you may won't be able to boot and you would need a locked CPU (i3,i5 or maybe i7-36x0) to remove the settings on the NVRAM if CMOS reset won't work.

* As Khenglish has noticed on another post its best to use as high BCLK OC as possible, since BCLK overclocking does not cause an automatic voltage increase, which multiplier increases do.

Less voltage means less heat/less power and maybe higher stability.

When i unlocked my CPU highest multi would be x40 for 1 core. When I tried to BCLK OC it would hang for everything above 102,3MHz. This means that with x40 for one core I would have a BCLK as low as 102,3MHz.

With 102,3MHz BCLK OC my 2 core Turbo would be 102,3 x 39 = 3,98GHz while my 4 core Turbo would be 102,3 x 38 = 3,89GHz.

When I lowered my 1 core Turbo @ x39 and kept 2-core turbo @x39 and 3/4-core turbo @ x38, I could again BCLK OC higher @ 104,7MHz. This means that with x39 for one core I would have a BCLK @104,7MHz.

With 104,7MHz BCLK OC my 2 core Turbo would be 104,7 x 39 = 4,08GHz while my 4 core Turbo would be 104,7 x 38 = 3,98GHz. Almost 100MHz higher. Got this?

*Yes, you can use the 99% CPU usage setting to prevent the overclock when not needing max performance or you can use Throttlestop to lower the multi or even use XTU to set your multiplier lower than the value you set in NVRAM. (keep in my mind that Throttlestop overrides the settings you have done in XTU).

We will be happy to attend your successfull Latitude OC. Feel free to ask anything else you need.

Mine has gone as low as 2.75GB with iGPU, dGPU, and eGPU on and allocated. Nando4 had problems getting his to lower though and it didn't do it until he changed his memory config. I think what you need to do is set to do a thorough boot every boot in BIOS. In nando's case it did not do a thorough boot until it detected a memory change.

It wouldn't go lower than 3.25GB. Maybe I needed to do a thorough boot. I did a CMOS reset(remove cmos battery for a few minutes) and everything is back to default (including CoreRatio and power limits btw). Anyway manual setting was easier and more fun. :) I may try if this work for me too when I have time.

@timohour, can i see a picture of your x2 2.0 setup? thank you!

...and what do you think is worth it for the cpu, 3630qm or 3720qm? TIA

This was a unstable setup that I used with a mini pcie to pcie extender and soldered some cables on a second mini pcie that was modified to check the ODD usage... I could send you a photo of the current setup.

Regarding the CPU I am quoting Tech Inferno Fan advice

The i7-37xxQM/i7-38xxQM better the i7-36xxQM with their +400Mhz locked turbo bins which some bios/uefi vars unlock. The i7-39xxXM have fully unlockable multipliers but you'll be TDP throttled and temp throttled to much lower multis than they are capable unless can increase TDP and improve cooling.

Hence why the i7-37xxQMs might be a good one to look at. Would suggest skipping the i7-38xxQMs. They get a bit more cache than the i7-37xxQMs. Hardly worth paying a premium for...

A i7-37xxQM/i7-38xxQM with unlocked turbo bins being the sweet spot for maximum performance at a reasonable price on these systems.

In simple words with the i7-3630QM your max turbo for 4-core would be 3.2GHz while for the 3720QM your max 4-core turbo would be 3.8GHz.

Also in my case the memory controller in my i7-3720QM is much better than the one in the 3630QM, since it would be rock stable @2234MHz dual channel memory while the 3630QM would crash even @ 2216MHz.

Just for your instance 3DMark06 CPU score difference would be almost 1100 points between a OCed 3630QM and an OCed 3720QM.

  • Thumbs Up 2
Link to comment
Share on other sites

@timohour, the "AHCI Link Power management" enabling with HIPM+DIPM makes a significant difference if your load the Intel Rapid Storage Tech (RST) driver. I have seen at least a further 0.7W idle drop over the MS driver when doing this. I'm pretty sure you'll see better than your current 7.2W min if trying this. I consider this a MUST DO when tweaking for battery life gains.

Updated driver with the latest intel, still no change, tried enabling/disabling the "AHCI Link Power management". I am guessing it is the msata to sata adapter that's causing the problem. Maybe I will try that with an HDD.

When I compared idle power consumption between a i7-3740QM and i5-3360M there was very little difference between the two. At most 0.2W.

No change with the i3 either. I can also tell that average power consumption has elevated a little.

OH.. and with ASPM tweak, please check if 'powercfg /energy' reports ASPM is enabled when you do the UEFI var mods as detailed at http://forum.techinferno.com/hp-business-class-notebooks/2537-12-5-hp-elitebook-2570p-owners-lounge.html#post38399 . I found enabling L0s L1 per PCIe port having little to zero affect. It was only a by enabling global ASPM support via a ACPI FACP bit enable that saw almost 2W gain on a 2570P.

I enabled global ASPM support using variables noted here. I think. I have to reenable now due to CMOS reset. Let me check

You are right "PCI Express Active-State Power Management (ASPM) has been disabled due to a known incompatibility with the hardware in this computer."

  • Thumbs Up 1
Link to comment
Share on other sites

We will be happy to attend your successfull Latitude OC. Feel free to ask anything else you need.

Thank you so much! This was very helpful indeed..

My CPU isn't an OEM version and throttles at x33 during my Throttlestop benchmarking. Should I go for the x42 (4 bins above my CPU's theoretical max), or x37 (4 bins above my CPU's real world max)? This was done on my HP 2570p.

Link to comment
Share on other sites

Thank you so much! This was very helpful indeed..

My CPU isn't an OEM version and throttles at x33 during my Throttlestop benchmarking. Should I go for the x42 (4 bins above my CPU's theoretical max), or x37 (4 bins above my CPU's real world max)? This was done on my HP 2570p.

In your HP 2570p your CPU was TDP throttled since the TDP limit for the HP 2570p is 40W.

On the 6430 the long-term TDP limit should be 45W.

For anything from 45 to 56.25W(mine has gone up to 52W) you would have the short-term limit and the CPU would TDP throttle after ~6 minutes of 100% utilisation.

To check your highest turbo you can check with the CPU-z program or XTU

UeDq4djm.pngrOFs5Mst.png

on those two pictures you can see that the highest available multiplier for the i7-3720QM is x36. +4 the highest would be x40.

EDIT: Use Throttlestop to check your highest multi. I think it is the easiest way...

An OEM i7-3840QM highest turbo would be x38 and +4 for the highest multiplier it would be x42.

Upgrademonkey here though claims the QCF1 QS cpu to be 3.7GHz so I suppose that your highest multi would be x37.

The best way to find out would be to make the swap and make the TDP table to see how far it can go stock...

Link to comment
Share on other sites

Once again, thank you @timohour. I didn't know the 35W limit was the factor holding back the 2570P. This makes perfect sense now.

Just to be clear on instructions:

  • Run Throttlestop, check max Multiplier achieved
  • Edit UEFI variables by adding 4 bin to the max multiplier (If upgrademonkey is correct, then it's x37 -> x41 1-C, x40 2-C, x39 4-C)
  • Reboot
  • XTU BCLK OC (A bit of Trial and Error, but is 104.7Mhz @ x40 a good starting point. I wouldn't know how to reverse the bricking/BSOD)
  • Reboot

Hope to get this OC working soon.

Link to comment
Share on other sites

Caution: Avoid using this mod as it may cause corruptions on your computer BIOS needing BIOS reflash or even motherboard replacement. Proceed at your own risk.

For unlocking your ME FW region better follow the unlock FlashDescryptor mod described here and flash modified ME FW.

How to unlock your ME FW region of your BIOS on a Dell Latitude E6430 (dGPU/iGPU) to flash a modified ME FW.

I was looking into my IFR for any variable that would help me enable ASPM and I stumbled across this:

Setting: Me FW Image Re-Flash, Variable: 0x228 
Option: Disabled, Value: 0x0
Option: Enabled, Value: 0x1

I made sure that I loaded the stock BIOS giving error 26 when tried to take a dump

fptw64.exe -d [I]name[/I].bin

rmLLLAbm.png

and then following Atonus' instructions to modify the UEFI variables using a bootable USB stick containing grub I set Me FW Image Re-Flash (variable 0x228) to Enabled (0x1).

Shutdown and then boot up and try again

fptw64.exe -d [I]name[/I].bin

utCd3MLm.png

It works! I could grab a full bios image.

On the following screenshot of the dump I got you can see highlighted the bits which show that the permissions for the CPU/BIOS and ME FW are not FF which means that the bios is still locked!!! (This makes sense since I didn't change anything in the BIOS).

xlOUV6Nm.png

Keep in mind that

* This mod is recommended for those who want to use a modified ME FW in order to BCLK OC using XTU.

* It won't help you unlock your Flash Descryptor though to set port 1 @ x2.2 (8Gbps). You need to use this mod.



NOTE1: It has a watchhog that change it again to 0x00 after a sleep cycle or shutdown but who cares. You can mod you descriptor to be definetely unlocked and enjoy!!!!

Too bad that's just for BCLK OC since I can only flash the ME Firmware region.

yHb1bcem.png

NOTE2: It seems like the Descriptor region is still locked. You will most probably still need the pinmod to change the descriptor settings (x2 2.0 & permanent unlock etc) but I can flash and reflash OC_ME fw without issues.

NOTE3:There is a MDES BIOS Status Code which is under the Platform Controller Hub (PCH) Firm Ware (FW) Configuration but I am not sure if I should mess with that...

NxpNQx2.png

MDES BIOS Status Code seems to be some debugging tool (replaced on newer systems with Intel PDA).

Setting this to enabled would hang the system needing an CMOS reset by removing every power source and the laptop CMOS battery for ~ 10-20 sec in order to clear any NVRAM & BIOS change and revert it to its default.

Image from iGPU model is property of

notebookcheck.netTFw7VuT.jpg

  • Thumbs Up 3
Link to comment
Share on other sites

Keep in mind that

* This mod is recommended for those who want to use a modified ME FW in order to BCLK OC using XTU.

* It won't help you unlock your Flash Descryptor though to set port 1 @ x2.2 (8Gbps). You need to use this mod.



NOTE1: It has a watchhog that change it again to 0x00 after a sleep cycle or shutdown but who cares. You can mod you descriptor to be definetely unlocked and enjoy!!!!

Too bad that's just for BCLK OC since I can only flash the ME Firmware region.

yHb1bcem.png

NOTE2: It seems like the Descriptor region is still locked. You will most probably still need the pinmod to change the descriptor settings (x2 2.0 & permanent unlock etc) but I can flash and reflash OC_ME fw without issues.

Ok I understand that the Flash Descriptor is for Port combinations (x2.2) and requires a Pin Mod, while flashing a modified ME Firmware allows for BCLK OC c/o XTU slider.

Could you please help clarify a few things for me?

  1. This UEFI variable 0x228 unlocks the ability to flash a (modified) ME firmware over the current one, right? But it undoes the command/re-locks itself after a sleep cyle/reboot.
  2. When UEFI variable 0x228 reverts back to locked/disabled, do we get to keep the newly flashed ME and its OC settings, or do we have to re-flash ME every boot?
  3. In the event I set the BCLK OC too high, what bad will happen and how do I troubleshoot/revert the changes?

Unlike the +400Mhz tweak, this BCLK OC sounds like it has a greater risk involved so forgive me for erring on the side of caution here.

Link to comment
Share on other sites

Ok I understand that the Flash Descriptor is for Port combinations (x2.2) and requires a Pin Mod, while flashing a modified ME Firmware allows for BCLK OC c/o XTU slider.

Could you please help clarify a few things for me?

  1. This UEFI variable 0x228 unlocks the ability to flash a (modified) ME firmware over the current one, right? But it undoes the command/re-locks itself after a sleep cyle/reboot.
  2. When UEFI variable 0x228 reverts back to locked/disabled, do we get to keep the newly flashed ME and its OC settings, or do we have to re-flash ME every boot?
  3. In the event I set the BCLK OC too high, what bad will happen and how do I troubleshoot/revert the changes?

Unlike the +400Mhz tweak, this BCLK OC sounds like it has a greater risk involved so forgive me for erring on the side of caution here.

1/2. UEFI variable temporarily give you the opportunity to update your ME FW with one that allows OC. Once you flash the modified ME FW you don't mind if it relocks unless. you want to return to the stock ME FW. (no obvious reason unless you want to return the machine for warranty reasons)

3. You can set the BCLK live with XTU. You don't have to set the BCLK in BIOS. In case you set a BCLK OC too high, it will BSOD and when you restart the watchhog in XTU will return your BCLK to 100MHz.

No risks involved more than those involved in every OCing...

  • Thumbs Up 1
Link to comment
Share on other sites

Thank you @timohour! At least OC can be used on the spot rather than from Bios/Boot. Waiting on the PE4C v2.1, and excited to get everything up and running.

Regarding the RAM OC (via XMP Profile), is this safe/stable to perform on top of a CPU OC (Multiplier + BCLK)? Just curious cause your post did mention benchmarking RAM on a 3720 instead of a 3630 due to workload/stress.

Link to comment
Share on other sites

  • 3 weeks later...

I've already ordered my E6430 with 1600*900 screen (just because of this thread), can't wait to get it. But its kinda hard having to say goodbye to my 2570p.

I don't think I will see a huge performance increase as I "only" have a 3630QM cpu. But I'm hoping the E6430 cooling will handle the quad core cpu better as it supports 45w TDP without problems.

Any chance of someone posting a guide on how to build your own eGpu adapter that supports x2 2.0 and works well with the E6430, because I already own two EXP GDC and a PE4C v2.1 is a bit expensive in Europe.

Link to comment
Share on other sites

I've already ordered my E6430 with 1600*900 screen (just because of this thread), can't wait to get it. But its kinda hard having to say goodbye to my 2570p. I don't think I will see a huge performance increase as I "only" have a 3630QM cpu. But I'm hoping the E6430 cooling will handle the quad core cpu better as it supports 45w TDP without problems.

Is this a dGPU model or an iGPU? Because heatsink & fans are different. In any case I think that your 3630QM would work great without temperature throttling issues.

Flashing a modified ME FW and applying ~5% BCLK OC as described here would elevate your 3630QM to an i7-3720QM level (up to 6% performance gain in 3dMark06 CPU score). You will already have to unlock your descriptor in order to set port 1 to x2 2.0 mode.

I would recommend though if you are planning to keep this system, that you side upgrade to an 3720QM/3740QM which would clock respectively up to 4C 3,98GHz/4,08GHz. Since a 3720/3740QM would start as low as 160€ on ebay this would be an 30-60€ upgrade which would give you an almost 20% performance gain on top of the stock 3630QM and on par or even better with high end i7s like the 4940MX.

Otherwise you better save money for Intel Skylake and upcoming ThunderBolt 3 32Gbps (x4 3.0)

Any chance of someone posting a guide on how to build your own eGpu adapter that supports x2 2.0 and works well with the E6430, because I already own two EXP GDC and a PE4C v2.1 is a bit expensive in Europe.

Building an custom eGPU adapter would require soldering skills and maybe won't work stable enough.

On the other hand using a PE4C along with 2 x PM100C would cost you 103$ (~93€) but would be a stable solution for x2 2.0 gaming.

Those two EXP GDC adapters connect to EC or mPCIe?

Link to comment
Share on other sites

Hi timohour,

Thanks for your reply.

Its a iGPU model.

I will probably keep this system for two or tree years.

I don't think I will be able to afford a new system anytime soon.

I have a good desktop for gaming. Its not the latest thing but Its still very capable. I7 990x, Rampage 3 Gene and 12GB DDR3 1600.

I get a kick at playing around with this eGPU stuff, I like it better than actually playing games. :D

Soldering skills wouldn't be a problem as they are a requirement at my current job and I need to practice.

If I manage to sell my 3630QM locally I will try to buy a 3720QM or higher.

Would a 3920XM make any sense in a E6430 or is the max recommend cpu a 3840QM due to its lower TDP?

I have a EXP GDC v.6 and a Beast, both of them EC. The Beast can be turned into mPCIe. The cable is reasonably cheap and easy to get. Would that help in any way?

Thanks for your input.

  • Thumbs Up 2
Link to comment
Share on other sites

Its a iGPU model.I will probably keep this system for two or tree years.I don't think I will be able to afford a new system anytime soon...

Soldering skills wouldn't be a problem as they are a requirement at my current job and I need to practice.

I have a EXP GDC v.6 and a Beast, both of them EC. The Beast can be turned into mPCIe. The cable is reasonably cheap and easy to get. Would that help in any way?

As noted on previews posts you would need two of the internal mini pcie ports to run @ x2 2.0. This means that you won't be able to use your EC gear.

I don't know the structure of the EXP GDC v.6 or the Beast (I never bought GDC products), but I guess that you would be able to solder the 2nd lane's pcie signals from your gear to a mini pcie adapter. Check this out or if possible upload a close photo of the bottom of your GDC product. (where we can see if there is a possiblity to solder cables)

In case this isn't feasible you would be able to use a riser but there would be a toll on signal's integrity.

If I manage to sell my 3630QM locally I will try to buy a 3720QM or higher.

Would a 3920XM make any sense in a E6430 or is the max recommend cpu a 3840QM due to its lower TDP?

Thanks for your input.

I am convinced that an XM CPU isn't going to make any huge difference in my laptop due to thermal limits (especially in your iGPU model that has a single copper pipe) but you can check the http://forum.techinferno.com/dell-latitude-vostro-precision/8176-%5Bguide%5D-dell-e6530-cpu-tdp-multi-unlocking-2.html#post135563 thread for more info/opinions.

In every circumstance a high binned CPU is appreciated, but I think that an XM cpu would need serious cooling mods on the cooler in order to perform as intended. I also feel that 4C @ 4GHz is more than enough.

Link to comment
Share on other sites

OCing Intel HD4000 on a Dell Latitude E6430 

After Khenglish posted a while back

Have you tried overclocking the iGPU with the IFR mods?
I promised that I would try that.

But is it worth it? Let's find out.
hanTppz.gif

I followed Atonus' instructions to modify the UEFI variables below using a bootable USB stick containing grub.

Setting: GT OverClocking Support, Variable: 0x16F
Option: Disabled, Value: 0x0 
Option: Enabled, Value: 0x1 
End of Options 

Numeric: GT OverClocking Frequency, Variable: 0x170
Default: 8 Bit, Value: 0x16
End 

Numeric: GT OverClocking Voltage , Variable: 0x171 
Default: 8 Bit, Value: 0x0 
End


Variable 0x16f should be enabled in order for the other two to be valid. So you set this to Enabled.

The other two are 8-bit numeric meaning that you can enter any value from 0-255 in hexadecimal so respectively from 0x00-0xFF.

On the GT Overclocking Frequency I realised that the resulting frequency was the result of [the decimal value] x 50MHz. e.g if I set value to 0x22 it would be 34 x 50MHz = 1700MHz . Anything under the maximum multiplier of your iGPU is not changing anything.

On the GT Overclocking Voltage I am guessing that there is a 0.01 increment for every value from 0x00 to 0xFF. E.g. I think that if I set the value to 0x05 I would have a 0,05V increment etc.

Since I didn't want any load from the CPU because a huge load on the CPU would trigger TDP throttling I used the Furmark 720p benchmark to test the performance gain from one frequency to the other. All tests where done using my i7-3720QM @ stock with 2x4GB Kingston RAM @ 2133MHz (except for the first which was done with i3-3120M and 2x4GB @1600MHz just for reference)

I didn't use BCLK OC but I hope I will have time to test further. I started from 1,25 GHz and went all the way up to 1,65GHz with a total performance gain (according to Furmark Benchmark) ~40%.

UPDATE: Using an adapter with an invalid ID I managed to set a higher TDP value (80W).

The maximum GPU multiplier that would work is x33 (1650MHz). It won't go higher than 1650MHz even if I set x34 and supply more Voltage and I don't think it is TDP or temperature throttled.

The following table shows the results:

Furmark Benchmark
 
GT Overclocking Frequency
Value
Resulting Frequency
(from GPU-z log)
GT Overclocking Voltage
Value
Voltage Increment
(speculation)
Memory Frequency
(Dual Channel)
Highest GPU Temperature
(from GPU-z log)
Highest GPU Power
(from GPU-z log)
Furmark 720p
Bencmark
unchanged 1100 MHz unchanged +0.00 V 1600 MHz   18.1 W 371
unchanged 1250 MHz unchanged +0.00 V 2133 MHz   17.9 W 515
0x1a 1300 MHz unchanged +0.00 V 2133 MHz   19.5 W 517
0x1b 1350 MHz unchanged +0.00 V 2133 MHz   21.1 W 532
0x1c 1400 MHz unchanged +0.00 V 2133 MHz 81 C 22.2 W 549
0x1d 1450 MHz 0x05 +0.05 V 2133 MHz 83 C 23.5 W 553
0x1e 1500 MHz 0x15 +0.21 V 2133 MHz 84 C 27.5 W 563
0x1f 1550 MHz 0x25 +0.37 V 2133 MHz 87 C 31.6 W 589
0x20 1600 MHz 0x40 +0.65 V 2133 MHz 93 C 37.4 W 639
0x21 1650 MHz 0x50 +0.80 V 2133 MHz 102 C   717



3DMark06 Benchmark
 

GPU Core Clock GPU Memory Clock 3DMark06 Score SM 2.0 Score HDR/SM 3.0 Score CPU Score Performance Gain
1250 MHz 2133 MHz 7850 2436 3250 7698  
1600 MHz 2133 MHz 9250 2942 3850 7516 up to 21%
Referencing Khenglish's results with [email protected] and iGPU 1.8GHz core 2250MHz memory on a Clevo P150EM
1800 MHz 2250 MHz 9647 3082 3952 8597 up to 25%

Intel HD4000, i7-3720QM @ 3.8GHz 4C, 2x4GB @2133MHz

Perforance Gain from stock clocks is almost 20% on 3DMark06 and possibly on most games.

Posted also Khenglish's results on a much more powerful Clevo to show that there is minor performance difference from 1,6GHz to 1,8GHz (most probably caused by higher RAM frequency as shown here) most probably because memory bandwidth starts to become an issue there.

On the initial 3DMark06 scene I noticed that fps went from 35 to ~43-45 which means that with this mod could probably change the state of a game from unplayable to playable, for gaming on the go with iGPU models.

Maybe I should consider a head to head comparison with the NVS5200M on the dGPU model too...

  • Thumbs Up 4
Link to comment
Share on other sites

Hi! Thank you for answering! I'm going to keep my hardaware, cause in dont have more money to spend.

You've told me that i have 3 options:

  • Use setup 1.x,
  • Use device manager to disable your dGPU PCI Express Port in Windows.
  • Use UEFI variables to manually lower your TOLUD to around 2 GB

If I have to buy the setup, I will, but how can I lower my TOLUD to 2 GB?

I've already disabled my PCI Express port but didnt work!

Thanks Again!

Link to comment
Share on other sites

Quote

Hi! Thank you for answering! I'm going to keep my hardaware, cause in dont have more money to spend. You've told me that i have 3 options:

  • Use setup 1.x,
  • Use device manager to disable your dGPU PCI Express Port in Windows.
  • Use UEFI variables to manually lower your TOLUD to around 2 GB


If I have to buy the setup, I will, but how can I lower my TOLUD to 2 GB?

I've already disabled my PCI Express port but didnt work!

Thanks Again!



Sorry for the late reply. Stupid notifications...:semi-twins:

What do you mean it doesn't work? What card/hardware/OS are you using? What error do you see in Device Manager?

It should work without modifications out of the box. Pls post your configuration and the exact problem (if you know)

EDIT found your first post (a few pages back) [FONT=tahoma] [/FONT]
Quote

[FONT=tahoma]I bought a few months ago a [/FONT][FONT=tahoma]pe4h, and recently a Asus Strix gtx 960 4 GB. [/FONT]



The problem is probably your PE4H which is ok for pcie x1 1.0...

I suggested you use setup 1.x or UEFI variables to change your port from gen 2 (x1 2.0) to gen 1 (x1 1.0)

Could you elaborate on your PE4H hardware? EC, mpcie ???

UEFI variables for Port 3 (ExpressCard) for pcie speed would be
Setting: PCIe Speed, Variable: 0xB4
Option: Auto, Value: 0x0
Option: Gen1, Value: 0x1
Option: Gen2, Value: 0x2
End of Options 
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use. We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.