CPU L2 Cache Memory setting in Vista

How to Set Vista for the CPU L2 Cache Memory Size​

   Information

The SecondLevelDataCache records the size of the processor cache, also known as the secondary or L2 cache. If the value of this entry is 0, Vista will attempt to retrieve the L2 cache size from the Hardware Abstraction Layer (HAL) for the platform. If Vista fails for some reason to retrieve the L2 cache size from HAL, then it uses a default CPU L2 cache size of only 256 KB. If the default value of the SecondLevelDataCache value is not 0, Vista will use the value (amount of memory) you set instead of 0 (for 256 KB) as the L2 cache size of. This value is designed as a secondary source of cache size information for Vista when the HAL cannot detect the L2 cache from the CPU.

   Note

This sets the SecondLevelDataCache inside the registry.

From Microsoft KB183063:

This is not related to the hardware; it is only useful for computers with direct-mapped L2 caches. Pentium II and later processors do not have direct-mapped L2 caches. SecondLevelDataCache can increase performance by approximately 2 percent in certain cases for older computers with ample CPU L2 memory (more than 64 MB) by scattering physical pages better in the address space so there are not so many L2 cache collisions. Setting SecondLevelDataCache to 256 KB rather than 2 MB (when the CPU has a 2 MB L2 cache) would probably have about a 0.4 percent performance penalty.

   Warning

Only set the value to what your processor actually has for the L2 cache size. Do not use a larger size than it actually has.

STEP ONE

Find Out the L2 Cache Size Reported by HAL

NOTE: This will show you how to see what the L2 cache size is for your CPU as it is reported to Vista by HAL. This value will be used in STEP TWO below to set the SecondLevelDataCache value for this as a backup in case HAL rarely could not detect the L2 cache from the CPU.

1. Generate a System Health Report.​

​

2. Click on the Hardware Configuration arrow on the far right to expand it. (See screenshot below step 3)​

​

3. Click on the Devices arrow on the far right to expand it. (See screenshot below)​

​

​

4. Under Devices, scroll down to Motherboard Classes. (See screenshot below step 5)​

A) Click on the + sign to expand​

root\cimv2:SELECT * FROM Win32_Processor.​

​

B) Click on the + sign to expand​

Win32_Processor.DeviceID="CPU0".​

​

5. In that list, you will see L2CacheSize and the value beside it for it's size to use it in step 6 in STEP TWO below.​

NOTE: For this example you will see 6144 to be used for my CPU's 2x6144 (12MB) L2 size. Now, here's a kicker. The newer Yorkfield Core 2 Quads CPUs use the entire cache for both cores unlike the older CPUs that do not share. So a 2x6144 (12mb) L2 cache would use 12288 (2x6144) instead despite what is reported in Vista for HAL.​

​

​

STEP TWO

Set the L2 Cache Size

NOTE: This will set the SecondLevelDataCache value in the registry, from the value gotten in STEP ONE, as a backup in case HAL rarely could not detect the L2 cache from the CPU and prevent having only a default L2 cache size of 256 KB used as L2 cache size instead by Vista.

1. Open the Start Menu.​

​

2. In the white line (Start Search) area, type regedit and press Enter.​

​

3. Click on Continue in the UAC prompt.​

​

4. In regedit, go to: (See screenshot below)​

​

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management

5. In the right pane, right click SecondLevelDataCache and click on Modify.​

​

​

6. Dot Decimal, and type in the value in bold (see table below) for your CPU L2 size from STEP ONE above. (See screenshot below table)​

NOTE: I have a CPU with a 2x6144 (12MB) L2 cache size, so I would use 6144 (Dec) or 1800 (Hex) since it is 2x. See STEP ONE above for how to get the L2 size for your CPU. Now, here's a kicker. The newer Yorkfield Core 2 Quads CPUs use the entire cache for both cores unlike the older CPUs that do not share. So a 2x6144 (12mb) L2 cache would use 122488 (Decimal) or 3000 (hex) instead despite what is reported in Vista for HAL.​

​

​

[TABLE=class:-grid,-width:-400][TR][TD]
Decimal Value
[/TD]
[TD]
Hexadecimal Value
[/TD][/TR]
[TR][TD]
256 KB
[/TD]
[TD]
100 (hex)
[/TD][/TR]
[TR][TD]
512 KB
[/TD]
[TD]
200 (hex)
[/TD][/TR]
[TR][TD]
1024 KB (1MB)
[/TD]
[TD]
400 (hex)
[/TD][/TR]
[TR][TD]
2048 KB (2MB)
[/TD]
[TD]
800 (hex)
[/TD][/TR]
[TR][TD]
3072 KB (3MB)
[/TD]
[TD]
c00 (hex)
[/TD][/TR]
[TR][TD]
4096 KB (4MB)
[/TD]
[TD]
1000 (hex)
[/TD][/TR]
[TR][TD]
6144 KB (6MB)
[/TD]
[TD]
1800 (hex)
[/TD][/TR]
[TR][TD]
8192 KB (8MB)
[/TD]
[TD]
2000 (hex)
[/TD][/TR]
[TR][TD]
12288 KB (12MB)
[/TD]
[TD]
3000 (hex)
[/TD][/TR]
[TR][TD]
16384 KB (16MB)
[/TD]
[TD]
4000 (hex)
[/TD][/TR][/TABLE]

​

​

7. Click on OK to apply.​

​

8. Your registry will look like this below for a 2x6144 (12MB) L2 cache memory size.​

NOTE: See STEP ONE above to find out your CPU L2 cache size.​

​

​

9. Close regedit.​

​

10. Restart the computer to apply the change.​

​

That's it,
Shawn

Related Tutorials

---

• How to Make a Program or Process Run with a Higher or Lower CPU Priority in Vista
• How to Set the Processor Affinity for a Application in Vista
• How to Change the Number of Processors Used at Boot Up in Vista

 

[Brink]

Hi Sale666,

Yep, as Johngalt posted, you would use 2048 KB (2MB)----800 (hex) in step 6.

Shawn

 

[ltwally]

Hi New64Pooter,

Welcome to Vista Forums.

It will be for your 2 x 4096 L2 cache. This will be 8192 KB (8MB) or 2000 (hex).

Shawn

Shawn, I believe you mis-understand what is being expressed when Intel says 2 x 4096.

Basically, all but the most recent (Yorkfield) Quad-core Intel chips are basically a pair of dual-core chips glued together. Their L2 caches are completely seperated. Cores 0 and 1 access L2 cache A, and Cores 2 and 3 access L2 cache B. In no way can Core 0 directly access cache B. It might be helpful to think of a Quad-core machine as a dual-processor dual-core machine.

That registry setting is telling Windows how much L2 cache can be accessed per core. This is a per-processor setting! Not a global sum of the total L2 cache. And, when I say "per-processor," it's important to think of that in the old single-core usage of the term. In modern parlance, that registry value is for the amount of L2 accessible per core.

Thus, the proper value is the maximum amount of cache that each core can address, which, in this quad-core Intel example, is 4096 kbytes. (Even though 2 cores actually share that 4096 kbytes.) The correct registry setting is '1000'.

For AMD systems that report 2 x 512 kbytes, the correct setting would be '400,' as each core has its own completely seperate L2 cache.

The newest Intel quad-core chips have a unified L2 cache. For these processors, you would use that simple value, as all cores access the same L2 cache.

To my knowledge, AMD is still producing chips with seperate L2 caches. However, their latest chips have a unified L3 cache.. but there is no registry tweak for an L3 cache that I am aware of.

-ltwally

 

[Brink]

Thank you Itwally.

I updated the tutorial to use another method of finding out the correct L2 cache size to use as reported from HAL to make it easier for everyone.

Shawn

 

[ltwally]

No problem. I've always been curious if this value has any odd performance penalties for Core2Duo processors, what with the way the cores are actually sharing that L2 cache. Unfortunately, you'd probably have to find a Windows Kernel engineer to give you the answer.

 

[Brink]

Itwally,

Mine is reported as 2x6144 in CPU-Z and as 6144 in the System Health Report (from HAL). So it seems that even though my Core 2 Quad Q9450 says it has a 12MB L2, it actually is being reported as having 6144 per core by HAL. I have not seen any performance hit so far with the setting as 6144 vs 12288. It would be rare for Vista to not get the size from HAL and have to resort to this registry setting anyway though.

 

[ltwally]

Itwally,

Mine is reported as 2x6144 in CPU-Z and as 6144 in the System Health Report (from HAL). So it seems that even though my Core 2 Quad Q9450 says it has a 12MB L2, it actually is being reported as having 6144 per core by HAL. I have not seen any performance hit so far with the setting as 6144 vs 12288. It would be rare for Vista to not get the size from HAL and have to resort to this registry setting anyway though.

What the HAL is (correctly) detecting is that you have 6megs of L2 cache per core. Now, the HAL is not intelligent enough to realize that cores 0 & 1 are actually sharing the same L2 cache, and ditto with cores 2 & 3.

Your processor looks something like this:

                FSB
                 |
                 |
--------------PROCESSOR--------------  
|                |                  |
|     -----------------------       |
|     |                     |       |
|   6meg L2               6meg L2   |
|     |                     |       |
|   -----------         ----------- |
|   |         |         |         | |
|  core0     core1     core2  core3 |
|                                   |
-------------------------------------

Whereas the classic model that was used at the time the Windows kernel was engineered looked a lot more like this

                                FSB
                                 |
            --------------------------------------------
            |                                           |  
-------PROCESSOR  0-------                 --------PROCESSOR  1------
|           |            |                 |            |           |
|         6meg L2        |                 |          6meg L2       |
|           |            |                 |            |           |
|          core0         |                 |           core0        |
--------------------------                 --------------------------

 

[yongkailoon]

I have a laptop with Core2Duo T9300 processor stated that it has 6MB L2 cache. From the tutorial, that processor is stated to have 12MB L2 cache but then it is set to 6MB L2 cache in Vista, may I know why is that ? Also since my T9300 has 6MB L2 cache, should I set it to 6144 then ? Thanks in advance.

 

[Brink]

Hi YongKailoon,

Welcome to Vista Forums. You will want to see what HAL is reporting in STEP ONE as the L2 size and use that size in STEP TWO.

My CPU basically has two separate dual cores (2x2 cores) with each dual core having 6MB. Together they have a total of 12MB, but HAL still registers it as 6MB being used per dual core. The top diagram posted by Itwally above can help make sense of this a bit more.

Hope this helps,
Shawn

 

[yongkailoon]

Hi YongKailoon,

Welcome to Vista Forums. You will want to see what HAL is reporting in STEP ONE as the L2 size and use that size in STEP TWO.

My CPU basically has two separate dual cores (2x2 cores) with each dual core having 6MB. Together they have a total of 12MB, but HAL still registers it as 6MB being used per dual core. The top diagram posted by Itwally above can help make sense of this a bit more.

Hope this helps,
Shawn

I see, from this I can actually assume that a standard dual core processor won't show 2 x of 3MB L2 cache since it is just one processor. From that diagram you mentioned, if I understand it properly the quad core is actually two processors of dual core so it makes it quad core. Anyway, I'll double check it with HAL but I hope my assumptions are right too. :D

 

[ltwally]

Hi YongKailoon,

Welcome to Vista Forums. You will want to see what HAL is reporting in STEP ONE as the L2 size and use that size in STEP TWO.

My CPU basically has two separate dual cores (2x2 cores) with each dual core having 6MB. Together they have a total of 12MB, but HAL still registers it as 6MB being used per dual core. The top diagram posted by Itwally above can help make sense of this a bit more.

Hope this helps,
Shawn

I see, from this I can actually assume that a standard dual core processor won't show 2 x of 3MB L2 cache since it is just one processor. From that diagram you mentioned, if I understand it properly the quad core is actually two processors of dual core so it makes it quad core. Anyway, I'll double check it with HAL but I hope my assumptions are right too. :D

You are exactly correct.

Now that you're no longer confused, I'll go ahead and confuse you a little.

The diagram above is valid for the Intel Quad-Core chips for the past 2 years or so. The very latest version of the Intel Core2Quad chips is code-named "Yorkfield," and looks like this:

                FSB
                 |
                 |
--------------PROCESSOR--------------  
|                |                  |
|   ---------12meg L2-------------  |
|   |         |         |        |  |
|   |         |         |        |  |
|   |         |         |        |  |
|   |         |         |        |  |
|  core0     core1     core2  core3 |
|                                   |
-------------------------------------

This is a much better design, as it allows all cores access to the same L2 cache - meaning they can trade data easily.

Now, to unconfuse you: the method that Brink uses will give you the correct value to use. So, just use it always.

 

[Brink]

LOL Itwally,

Thank you for clearing that up, or did you? LOL

Thank you,
Shawn

 

[SCSIraidGURU]

You could just get a pair of Intel quad core Xeons. They have a 12MB cache on each of them.

 

[ltwally]

You could just get a pair of Intel quad core Xeons. They have a 12MB cache on each of them.

Xeons are basically Core2Duo and Core2Quad processors. Sometimes the FSB speed is higher, or sometimes they have more cache - but the processor design is fundamentally the same.

Which means..... Xeons suffer from the confusion. Older generation of Quad-Core Xeons will have a seperated L2 cache, while the latest generation (built on the Yorkfield core) will have a single, shared L2 cache.

Of course, if you want to use a Xeon, you'd better make sure that your motherboard supports them! They (usually) use the same socket and pin-out as the desktop Core2Duo & Core2Quad, but if your system BIOS doesn't support the processor, you're up s**t-creek.

Honestly, the only real advantage to running Xeons is that they support multi-processor systems (ie. two quad-core chips on one motherboard). Unless you require that, just buy yourself a Core2Extreme - it'll probably cost you less than the equivalent Xeon.

 

[SCSIraidGURU]

My new God Box:supermicro X7DWA-N motherboardDual Intel E5430 quad core 2.66 GHz 1333 MHz FSB Xeons16 GB FB-DIMM DDR667 Ecc Reg LSI Logic MegaRAID 8708EM2 with 128 MB cache SAS RAID controller, PCIe x8Hot Spare: 1 x Seagate Savvio SAS 146GB 10.6K RAID 1: 2 x Seagate Savvio SAS 146GB 10.6K eVGA 8800 GTS 640MB PCIe x16 video cardSony DRX-830U USB2 DVD burnerCooler Master Stacker 810 PC Power and Cooling Turbo-Cool 1kw-SR 1000W power supply2 x 20A@120V double pole Independent circuitTripp Lite 2400W Line Conditioner for computer: Workstation Tripp Lite 2400W Line Conditioner for computer: Monitor, printers, Firewire drives, USBI do a lot of A/V and CAD work on this workstation. It was designed for Photoshop and other applications that can run on multiple cores

 

[ltwally]

You get points for your system, but loose them for putting your entire post into one long run-on sentence. Carriage returns. Use 'em!!

Seriously: wtf do you need that kind of hardware for? You've got the wrong video cards for 3d rendering... The system looks like one insanely powerful general-purpose workstation to me. Do you just have too much money and not know what to do with it? If so... let me help you with that problem...

How much did you blow on that system? $8,000? more?

 

[SCSIraidGURU]

Replacing my video card in a few months. A lot of money. I am waiting for Photoshop CS4 to be released and then deciding on a video card. I have one PCIe x16 slot. I was looking at the 8800 GX2.

 

[ltwally]

Replacing my video card in a few months. A lot of money. I am waiting for Photoshop CS4 to be released and then deciding on a video card. I have one PCIe x16 slot. I was looking at the 8800 GX2.

Very little of anything done by a video card has any impact on Photoshop, though. For the most part, one video card is just as good as another when it comes to that.

 

[SCSIraidGURU]

I am counting on 8 x 2.66 GHz cores, 16GB of RAM, and SAS RAID arrays to make up for graphics card. I don't do much 3D work anymore. Most of my CAD work is EE or architectual 2D design. My A/V is post processing on RAW images to JPG in Photoshop. Sony Vegas Movie Studio to make slide show video DVDs.

 

[eXXXodus]

Is there any way yet to change the L3 cache size? I have a 9850BE that has a 2MB L3 cache. Thanks.


_______________________________________________________________________
System specs:
AMD 9850BE - ASUS M3N-HT mobo - EVGA 8800GT Superclocked - OCZ Vendetta Cooler - Maxtor 500GB HDD - OCZ 700W GameXtreme PSU - 4GB OCZ Platinum 1066 DDR2 Memory - NZXT Apollo case w/ Sentry1 controller

 

[Brink]

Hi eXXXodus, and welcome to Vista Forums.

None that I am aware.

Sorry,
Shawn