Jelly Donut 2019 年 12 月 24 日 下午 6:30
9700k owners - What are your OC/Power Silicon Lottery results? (High Heat?)
So earlier this year I made a thread asking about the experience from 9900K/9900KF owners. As I was skeptical about heat since I am OC'ing and the country I live in is pretty warm all year round (Singapore) and considering I don't multitask heavily, I saved some money and went for a 9700k.

My 9700k had a pretty good silicon. Was able to hit 5.1Ghz, 4.9Ghz (AVX) @ 1.35V, LLC4 (I was able to do 5.2 and 5.0 @1.365V, but it was really difficult to cool). However, it consumes power like a 9900k. While doing an AVX workload, it consumes nearly 200W of package power as reported from the motherboard as well as got pretty toasty (hitting 90C during an AVX workload on prime 95 for e.g.). I use a Side mounted Corsair H150i Pro Liquid Cooler.

For those who overclock their 9700k here.. Is this normal heat and package power consumption? Because, I feel that 9700Ks are like 9900Ks tested by Intel that didn't qualify for the 9900K's specs, so they just disable the HT and sell as a 9700k (and it looks like a more inefficient silicon compared to the 9900K, let alone the 9900K Special Edition)

Let me know what are your experiences with this chip!

Specs, Motherboard and OC settings:
Motherboard: ASUS Maximus XI Gene
PSU: Cooler Master V750 (750W 80 Plus GOLD)
Cooler: Corsair H150i Pro (360mm AIO) w 6 fans: 3 Noctua NF-F12 iPPCs for Push and 3 Cooler Master RGB Fans for Pull

OC Settings:
Ai Overclock Tuner: XMP II
BCLK Frequency 100.0000
ASUS Multicore enhancement: Disabled
SVID Behaviour: Intel's Fail Safe
AVX Instruction Core Ratio Negative Offset: 2
CPU Core Ratio: Sync All Cores, 51
DRAM Frequency DDR4-3200Mhz
Xtreme Tweaking: Enabled
CPU SVID Support: Disabled

CPU Core/Cache Current Limit Max: 255.75 (MAX)
BCLK Aware Adaptive Voltage: Enabled
CPU Core/Cache Voltage: Manual Mode
CPU Core Voltage Override: 1.350
CPU Cache Ratio Min and Max: 43

-External Digi+ Power Control
CPU Load-Line Calibration: Level 4
CPU Current Capability: 140%

-Internal CPU Power Management: Turbo Mode Parameters-
Long Duration Package Power Limit: 4095 (MAX)
Package Power Time Windows: 127
Short Duration Package Power Limit: 4095 (MAX)

*All other settings not listed are set to their default settings, being it Auto or any other assigned by XMP II.
最后由 Jelly Donut 编辑于; 2019 年 12 月 24 日 下午 6:33
< >
正在显示第 16 - 23 条,共 23 条留言
tacoshy 2019 年 12 月 25 日 上午 11:38 
引用自 Snow
Either way Intel can't easily do stuff what TSMC can, that's for sure. Basically, if some new CPU design fails for Intel - it's gonna be a financial disaster, but if the same happens to TSMC - they'll just have an angry customer. I wonder if those "Mac going AMD" rumour is true, as if it is - it's yet another huge market lost for Intel, and without Intel around we'll have Zen2 refreshes for next 10 years lol.
Not really, TSMC is not better then Intel in manufacturing process. However AMD (AMD =/= TSMC) is more inovative atm then Intel.
However Intel doesnt really care and fails are not an financial issue as they have a complete different market. Intel also spend more on development then AMD entire revnue. AMD has a long history of flops which caused them to nearly disappear from the market, just their console chips kept them afloat befor Ryzen.

The thing that many ppl miss, what we see here or with gamers is only a small segment of the market that profit wise is to little to make a real impact. The real money lies in the server and professional market where Intel rules. Thats also why they keep having such high revenue even though they tooked a big hit on the consumer gamer market. ANd its still heavily increasing:

https://www.macrotrends.net/stocks/charts/INTC/intel/revenue
最后由 tacoshy 编辑于; 2019 年 12 月 25 日 上午 11:39
Snow 2019 年 12 月 25 日 上午 11:43 
引用自 tacoshy
The thing that many ppl miss, what we see here or with gamers is only a small segment of the market that profit wise is to little to make a real impact. The real money lies in the server and professional market where Intel rules.
Is there any way to find out exact or relative numbers on Intel's revenue sources? I.e. 10% from desktop CPUs, 3% from connectivity modules etc. As I don't deal with any sort of "big and expensive" hardware, it's no surprise I missed the elephant in the room, and I'd like to see the whole picture if that's possible.
tacoshy 2019 年 12 月 25 日 上午 11:49 
引用自 Snow
引用自 tacoshy
The thing that many ppl miss, what we see here or with gamers is only a small segment of the market that profit wise is to little to make a real impact. The real money lies in the server and professional market where Intel rules.
Is there any way to find out exact or relative numbers on Intel's revenue sources? I.e. 10% from desktop CPUs, 3% from connectivity modules etc. As I don't deal with any sort of "big and expensive" hardware, it's no surprise I missed the elephant in the room, and I'd like to see the whole picture if that's possible.

not really but if you think about the revenue between the different years and the impact with AMD market share from Ryzen this year, my best educated guess is that the consuemr market has 5-10% of their total revenue.

If you see a server with something like a Xeon Phi (large server centers for companies or universities have litterally hundret of those) which cost more then 3-4 Grand, or the Intel-X Series CPU's which cost easily trice or mroe then a consumer CPU's, you see where the real profit lies.

And frankly even though TR and Ryzen have a good core ammount they still perform worse with Adobe products and thigns that do more then just rendering. So Intel proberly will stay in the professional market ahead for a logn time unless AMD can get a breakthrough with Epyc finally.
Snow 2019 年 12 月 25 日 下午 12:08 
引用自 tacoshy
Intel-X Series
Which reminds me there were some recent-ish news about Intel cutting extreme CPU prices in half, just like that. Seems like competition is strong on that market as well these days.
引用自 tacoshy
And frankly even though TR and Ryzen have a good core ammount they still perform worse with Adobe products and thigns that do more then just rendering.
I've seen various articles about benchmarks somehow being biased towards Intel CPUs. like they use Intel hardware more efficient than they do use AMD. Is this Adobe thing the result of Adobe software being made to use more Intel's features, or does this happen because Intel CPUs are a better suit for jobs Adobe products put on them? If there's a way for us to know such things that is.
最后由 Snow 编辑于; 2019 年 12 月 25 日 下午 12:09
tacoshy 2019 年 12 月 25 日 下午 12:30 
引用自 Snow
I've seen various articles about benchmarks somehow being biased towards Intel CPUs. like they use Intel hardware more efficient than they do use AMD. Is this Adobe thing the result of Adobe software being made to use more Intel's features, or does this happen because Intel CPUs are a better suit for jobs Adobe products put on them? If there's a way for us to know such things that is.

its optimization and support espaiclly instruction sets (Threadripper misses AVX-512, it has AVX and AVX-2 but still missing the AVX-512). So yes it is kinda biased. they work hand in hand for a long time espacially because befor Ryzen AMD didnt had anything to really offer in that area.
Same as some games work way betetr with Nvidia hardware as the got the help from Nvidia to develop and get the knowledge how to sue the hardware more optimal.

yes TR and Ryzen amde an impact on the professional market aswell as it the core ammoutn for less moeny was a thing to consider and not everything is optimized toward Intel.
Also Epyc had some ncie advantages stills truggle in the scientific field where the Xeon Phi 7295 is the only "viable" option.
Snow 2019 年 12 月 25 日 下午 12:34 
Thanks, it's always nice to learn few things ♥
hawkeye 2019 年 12 月 25 日 下午 4:07 
The cost of cpus isn't a factor for most corporate buyers. Hardware costs can be tax depreciable. Annual software licences can cost more than hardware - the better the cpu the higher the cost also. Power usage and cooling costs are important. Swapping something that is known to work well for something whose performance is unknown is a risk that few IT platform managers would take for business critical systems. And it's company money not your own.
tacoshy 2019 年 12 月 25 日 下午 5:27 
Good points hawker at which point it might be worth to mention, that many large company doesn't bother to repair PC's, they simply replace them. Cost less time and headache as cost of the PC is not an issue but a downtime where somebody can't work efficiently.
< >
正在显示第 16 - 23 条,共 23 条留言
每页显示数: 1530 50

发帖日期: 2019 年 12 月 24 日 下午 6:30
回复数: 23