Weekend benching, just to share some of my OCed temp comparison results between Intel and its equivalent counterparts from AMD on the same test bench with the same cooling setup. Results are based on Prime95 custom blend 1344k for 30mins:



Cooling setup:
CPU water block: EK Supremacy EVO RGB
Radiators: 2x XSPC RX360
Pump: Swiftech MCP655
Reservoir: EK Multioption X2 Advanced
Tubing: XSPC 1/2" ID 3/4" OD white
Fans: 12x Scythe GT AP-15 1850rpm 120mm
Thermal paste: Thermal Grizzly Kryonaut
Case: DimasTech Easy Dual V2.5 Bench Table

8700K vs 2600X (6C/12T)
8700K (no delid) OC 5.0GHz @ 1.280v => max 77.0C
2600X OC 4.2GHz @ 1.4125v => max 69.5C

9900K vs 2700X (8C/16T)
9900K (no delid) Stock 4.7GHz @ 1.172v => max 63.0C
9900K (no delid) OC 5.0GHz @ 1.279v => max 81.0C
2700X OC 4.25GHz @ 1.456v => max 74.4C

Some gaming benchmarks...

FC5 game is a good example that demonstrates the CPU scaling. If you looked again at the chart above, at max OC the Ryzen CPUs are significantly slower than Intel especially on the minimum fps. In fact, stocked 8700K and 9900K are already faster than the overclocked Ryzen equivalent.

In terms of thermal performance, Ryzen are actually hotter clock-to-clock. If you observed closely at my temp comparison above, a 9900K at stock 4.7GHz is much cooler than an OCed 2700X at 4.2GHz. Same thing for 8700K vs 2600X.

Not to mention the higher voltage that the Ryzen needed at these clock frequencies. Now take a look at the power consumption as the following, you will notice that a stocked Intel consumes much less power than an OCed Ryzen, and Intel is faster.



This post has been edited by owikh84: May 20 2019, 07:53 AM

Good news guys, now Malaysians can buy almost anything from Newegg.



https://www.newegg.com/global/my-en/

For rare items that are not available locally, yes can buy from Amazon/Newegg etc if you don't mind the warranty hassles.

[Anand] Intel Announces 8 Core i9-9900KS: Every Core at 5.0 GHz, All The Time


Source: https://www.anandtech.com/show/14402/intel-...-core-turbo-cpu

My board on the sale LOL
Thinking to let the Gene go as well after getting back from RMA (one RGB LED dead).

Haven't send, need to arrange time crossing the Penang bridge in the weekdays.
Based on my previous RMA experience with this Penang Royap Club, 10days for motherboard, 7 days for graphics card.

Not dead, actually this LED just can't sync color with the rest of the LEDs. I suspect the electrical trace path on the board is broken.
Had seen the same thing with my other RGB products, so I think it's a common issue.

CPU Load-line Calibration (LLC) set to Level 8 (highest for ROG), this will give the least vdroop/vdrop.

This post has been edited by owikh84: Aug 8 2019, 10:46 AM

Not an apple to apple comparison bro.
My highend custom wc is always better than AIO by 15c.
On top of that, you have to take account of the variations in every chip, different vcore required, different mobo (VCCSA, VCCIO settings), uncore, memory speeds, chassis airflow etc. All these factors contribute to the heat as well.

This post has been edited by owikh84: Aug 10 2019, 10:35 PM

For my custom loop configuration, you gotta prepare at least RM3.3k.

Supremacy EVO RGB: RM369
EK D5 Vario: RM350 pump + RM200 pump top
EK-RES X2 250: RM200
2x RX360: 2x RM400?
12x GT AP15: 12x RM90
10x Comp Fitting: 10x RM30
Primoflex LRT Flexible Tubing 5ft: RM60
Others not counted: Cable sleeving for fans + pump wires, 2x 6-fan splitters, etc

Of course there are bunches of newer parts with ARGB illumination (more expensive) so just take this config as a references. Lol...

Taobao China.
Don't think Alphacool is available locally.

Here you go: 【欧酷冷排散热器全紫铜Alphacool NexXxoS XT45 120/240/360/480mm】https://m.tb.cn/h.eQTZROI?sm=295239 点击链接，再选择浏览器咑閞；或椱ァ製这段描述₴rDs9YQT2Xjv₴后到👉◇綯℡寳👈

Based on my personal experiences with more than 10 transactions and total value of each shipment over RM500, never been taxed for Tabao purchases.

Getting better there, B-die should be able to do 4k CL17-18, 4133 CL18-19 @ 1.35-1.40v etc.

I don't think you can adjust any RAM timing with MemTweakIt.
Anyway, first post updated with the latest version: https://drive.google.com/file/d/14PhC6UJjPE...EVAnKcLYfH/view


From 2400 to 3900, albeit timings on the high side, pretty sure you will see a performance boost as true latency has now been reduced.
Would be nicer if you're still able to reduce the timings with additional voltages.
Good luck.

Check this out guys, nice 9900KS review.

Had fun benching with 9900KS, not golden but significantly better than my last 9900K (5.0GHz @ 1.280v). I'm limited by thermal, custom loop probably needs some maintenance. Voltage/core temp scaling on RB v2.56 AVX for 1 hour:

Core/Cache (GHz)	vCore (v)	Min (°C)	Max (°C)	Stable RB 1H?
5.0/4.3 (Stock)	1.181v	32.9	77.4	Yes
5.0/4.7	1.181v	35.4	79.5	Yes
5.1/4.7	1.225v	36.1	85.8	Yes
5.2/4.7	1.270v	37.4	88.1	No

RB for 1H at stock clock:


Quick benchmarks at 5.2/4.7...

Cinebench R20:


Cinebench R15:


IMC test with B-die...
DDR4-4600 18-22-22-42-2T @ 1.425v BIOS set:

For GIGABYTE, that vCore shown in your CPU-Z is wrong because it's not read from a die-sense sensor like ASUS. Your actual vCore can be read from HWiNFO64 as VROUT.
Also, RealTemp can only read up to 4 cores (GT up to 6 cores).
I'd suggest using HWiNFO64 to monitor temps, voltages, CPU power package etc.
A Loadline Calibration (LLC) of High or Turbo is recommended for GIGABYTE to get better vDroop control.
You might want to go through first post of this thread for GIGABYTE OC guide.
Good luck

TS benchmark at 5.2/4.7/ DDR4-3200 CL16 vs 4266 CL19,

3200 CL16: CPU score 12.1k


4266 CL19: CPU score 12.5k


That's an 3.3% improvement with faster RAM speed.