My Experience Upgrading the AREA-51m R1 to an i9-9900KS with a Custom Optimized Profile

[Amir.s]

 

Time spy test score

Old Cpu  9900K:

https://www.3dmark.com/spy/56481843

~~~~~~~~~~~~~~~~~~

New Cpu 9900KS :

https://www.3dmark.com/spy/64018720

 

After upgrading my Alienware Area-51m R1 to the Intel Core i9-9900KS, I noticed that almost every user who performed this upgrade experienced significant CPU performance degradation. Most reported Time Spy CPU scores were only between 8000 and 9000 points due to power limits, EDP throttling, BIOS limitations, and the lack of a proper 9900KS profile in the BIOS.

I decided to fully optimize and tune the system manually. After testing multiple BIOS versions, tweaking hidden settings, optimizing memory timings, tuning power behavior, and creating a custom optimized profile for the 9900KS, I finally achieved a stable and fully optimized result.

Previously, with the standard i9-9900K, the best temperature I recorded during Time Spy was 96°C. However, after upgrading to the i9-9900KS, the system delivered noticeably higher and more stable performance while running significantly cooler, with a maximum Time Spy temperature of only 71°C.

My final Time Spy score reached 18,411 overall, with a CPU score of 11,206 — paired with an EVGA RTX 3080 Ti XC3 that scored 20,767 in graphics performance, which is currently the highest known CPU score achieved with an i9-9900KS on an Alienware Area-51m R1.

What makes this result even more impressive is that the CPU maintains excellent temperatures and efficiency while delivering desktop-class performance inside a laptop platform.

 

 

I compared the performance of my laptop using two processors, the 9900K and the 9900KS. You can see the temperature difference between both CPUs in the Time Spy benchmark. There is more than a 25°C temperature gap between these two processors on the Area-51M. The reason is the power limitation of the 9900KS. By undervolting with XTU, you can achieve higher performance at lower power consumption and significantly lower temperatures.

 

 

 

The 9900KS on the Area-51M is power-limited by the BIOS. Since voltage profiles for the 9900KS are not defined in the BIOS, it only allows the CPU to use up to 130W. I am trying to remove this BIOS-imposed limitation. However, even with this 130W limit, you can still achieve significantly lower temperatures compared to the 9900K or any other processor, along with higher performance. If I can remove this limitation, I can increase the CPU score in the Time Spy  cpu score benchmark up to 13,000.

 

 

System highlights:

• Alienware Area-51m R1 

• Intel Core i9-9900KS

• Custom optimized BIOS/RU tuning

• DDR4-2666 overclock to 2933 CL17memory tuning

• Fully optimized thermal and power behavior

• Stable high-performance operation

 

 

[2CPU]

Very nice work.

[Amir.s]

10 hours ago, 2CPU said:

Very nice work.

I’m waiting for you to improve your CPU score in Time Spy and send me a really good record so I can enjoy seeing the result.

 

 

If I had an RTX 5080 right now, I would definitely set an incredible benchmark record with this laptop. Unfortunately, my card is only a 3080 Ti. Come on, let’s make an impressive record with the 5080.

[2CPU]

I set the max for all cores to 47x in OC BIOS settings.

 

Then did this under-voltage:

 

 

[Amir.s]

On 6/19/2026 at 8:41 PM, 2CPU said:

I set the max for all cores to 47x in OC BIOS settings.

 

Then did this under-voltage:

 

 

Go into XTU and set the Core Voltage Offset to -0.120V, then set all cores to 4.8GHz and test it. Your CPU performance should improve significantly. If it’s stable, try lowering it further to -0.125V and test again. The CPU should stop throttling, and temperatures should drop as well.

 

 

Your CPU score is terrible. You should be getting somewhere between 10,200 and 10,400 points.

 

 

On my 9900K, I was completely stable at -0.140V with all cores running at 4.8GHz, and I had absolutely no throttling.

[2CPU]

Timespy Dev wrote that the 5080 causes CPU overhead issues due to their code. (so it is not ideal to test with)

 

Just tried 48x @ -0.120V and came in similar. Not sure if because I have steam also running since that's how I got the demo.

CPU at 85.6*

 

Fans are set to full speed.

[Amir.s]

17 hours ago, 2CPU said:

Timespy Dev wrote that the 5080 causes CPU overhead issues due to their code. (so it is not ideal to test with)

 

Just tried 48x @ -0.120V and came in similar. Not sure if because I have steam also running since that's how I got the demo.

CPU at 85.6*

 

Fans are set to full speed.

 

17 hours ago, 2CPU said:

Timespy Dev wrote that the 5080 causes CPU overhead issues due to their code. (so it is not ideal to test with)

 

Just tried 48x @ -0.120V and came in similar. Not sure if because I have steam also running since that's how I got the demo.

CPU at 85.6*

 

Fans are set to full speed.

It has nothing to do with the RTX 5080.

Install XTU version 7.14.2.93 and set Core Voltage Offset to -0.130V. Also set all CPU cores to 47x.

Check in Task Manager to see how much the processor is being utilized. If CPU usage is between 2% and 8%, that is normal.

Use HWiNFO to check the core temperatures. The CPU is not throttling.

A score of 7500 is definitely not normal at all.

[Amir.s]

4 minutes ago, Amir.s said:

 

It has nothing to do with the RTX 5080.

Install XTU version 7.14.2.93 and set Core Voltage Offset to -0.130V. Also set all CPU cores to 47x.

Check in Task Manager to see how much the processor is being utilized. If CPU usage is between 2% and 8%, that is normal.

Use HWiNFO to check the core temperatures. The CPU is not throttling.

A score of 7500 is definitely not normal at all.

Time Spy is still a valid benchmark for evaluating both GPU and CPU performance.

I am a professional overclocker and have extensive knowledge in this field. My benchmark results can be found on HWBOT.

On an Area-51m, your processor should normally achieve a CPU score between 10,000 and 10,500 in Time Spy, unless there is an issue affecting the system’s performance.

[ssj92]

1 hour ago, Amir.s said:

Time Spy is still a valid benchmark for evaluating both GPU and CPU performance.

I am a professional overclocker and have extensive knowledge in this field. My benchmark results can be found on HWBOT.

On an Area-51m, your processor should normally achieve a CPU score between 10,000 and 10,500 in Time Spy, unless there is an issue affecting the system’s performance.

Correct, this is my Area-51m result from 2019. Haven't tried recently but should go but now since I have 2933Mhz memory clock. 

 

 

https://www.3dmark.com/spy/7167581

 

 

[Amir.s]

4 hours ago, ssj92 said:

Correct, this is my Area-51m result from 2019. Haven't tried recently but should go but now since I have 2933Mhz memory clock. 

 

 

https://www.3dmark.com/spy/7167581

 

 

Nice score 😍

[Amir.s]

My score  intel XTU 2 benchmark

Area51m-R1

cpu 9900ks

freq : 5100 Mhz all core

 

Link:  https://hwbot.org/benchmarks/xtu_2.0/submissions/6029701