Here’s a 5minutes idle, averaged 0.58W:
for the record: is that the power draw on battery or from the wall?
What jumps to my attention:
Throttlestop is reporting C0% of ~6% on your system where my system achives 1% to 2%. Your CPU spends more time in the fully active C0% state - which is probably the reason for the higher power draw.
Some process is keeping you CPU occupied. I recommend you to experiment a bit with disabling third party services: e.g. disable google drive and displaycal and see how much C0% reduces.
Be aware: your ThrottleStop is currently only in monitoring mode. To apply ThrottleStop settings you have to click on ‘Turn On’
Is this CPU power from HWinfo64? System power from HWinfo64? Or measured some other way?
I could have been clearer about that:
The reported values are for CPU package power draw as shown by ThrottleStop.
Looks like Google Drive was pulling a little bit of CPU. DisplayCal was basically nothing. FxSound however was over a watt! That was quite surprising. I had that installed to make the speakers sound better, one of the biggest problems with the Framework laptop in my option.
1 Like
Thanks! In that case, my 12th gen, with a stock windows install (and framework driver package installed), is very similar. On AC power / balanced profile and idle it’s about 0.5-1W on both DC/AC power with balanced profile.
IDo these power settings affect performance? My cinebench came in at 3703. I’m very pleased with the fan noise reduction, but a ~25% drop in speed does not seem balanced.
My system is an i5-1135g7 with 32gb of ram running windows 11.
According to framework’s internal testing the score should be over 5000.
Did you cap the TDP, or was there some windows updates running in the background?
In general though, you can’t have both great battery life, AND, high performance. … Unless you have a larger battery.
I realise I can’t have my bits and eat them too. My goal was to keep my framework from sounding like a tiny vacuum during simple tasks.
After removing all the settings in this thread, disabling windows update in services.msc, and setting the power option to performance, my score went up to 4800. After installing bios 3.10, my score went up to 5003. That was especially surprising because I couldn’t find anything in the bios release notes that mentioned a performance improvement.
Something is still amiss as the lab benchmark is 5100.
I would consider this close enough.
The results will vary from machine to machine and from run to run.
3 Likes
If you are just using your Framework for general stuff, then just switch off the CPU turbo boost in the BIOS.
I switched mine off a month ago. Have noticed no difference in day to day usage and it’s quieter.
Only takes a minute to switch it back on if you need the boost.
2 Likes
I thought switching it off in BIOS only applies to the governor in the BIOS? Windows override the control when it boots.
Nope just switch off the boost in BIOS and my Framework locks to max 2.8Ghz under load in Windows 99.9% of the time.
2 Likes
I switched off boost in my BIOS and it appears to be working. I don’t use monitoring software but the clock speeds look correct when I open task manager and the laptop is dead silent for the most part.
1 Like
Yeah I think due to tech reviews and benchmarks we get fixated that we need faster and faster peak performance but day to day? For 90% of what we do? Nope, it just has to keep up with what we are doing. Modern chips have plenty of head room nowadays. I bet I could downclock (if I could) this Framework to a max of 2GHz and I still wouldnt notice in most of the stuff I do. I got fast dual channel CL20 ram and a SSD that can do 6900+MBps which all helps keep stuff smooth.
1 Like
@Jason_Dagless are you noticing any battery life improvements after turning off turbo? I might do the same when my framework arrives. Great idea and thank you for sharing
@seadub I dont own a Framework, but I do own a Latitude 9520 with an i7-1185G7.
Disabling turbo keeps temps really low (idling at 40C and at most, it goes up to 47C-50C, on a room that’s at 27C-30C) and battery life is better because there’s not a spike in power usage.
I dont know what’s the Package Power on the Framework Laptop but on here, it’s set at 31W according to the Intel Power Gadget (HWInfo64 says its 28W), so with Turbo ON, Windows will jump up to 4GHz and spike up to 20W whenever it feels it needs that extra power.
Without Turbo, at most, you will a power usage of 7W-8W. These are the figures reported by the Power Gadget.
Yeah I’ve watched HWmonitor when testing and with boost as mentioned above, the temps are way lower, fan usage is almost nil for web and general use. Wattage was 30W peak with boost and well under 10 with it off (around 4W at idle). Just makes the laptop feel more “grown up” when its not huffing and puffing all the time. I really don’t know what Intel are thinking these days if they feel thats acceptable behaviour for a product. Imagine 8 or 9 such laptops in a meeting room…
One tip I would say is to make the rear foot strip at the back maybe 1mm deeper to give more clearance for air flow. Shame Framework used a strip rather than feet, as its easier to stick something to an existing round or square foot.
On battery running WPrime with no boost for my 1165G7 Max wattage is 18.3W, Max temp 69C and Max CPU was 2811MHz. The fan does come on but it sounds like its running at maybe 33%. You can just hear it.
1 Like
Yeah but its nicer to add to a larger existing foot surface area than add something new elsewhere. You keep the existing support area too rather than expanding or reducing. When I de-commision old laptops I keep all the rubberwear from them cos if I get a nice laptop in I can hand on, often the rubberwear is missing. My rubber laptop feet collection is extensive! lol
1 Like
The ‘strip’ acts as a barrier to isolate the hot air from being pull back in underneath due to air flow / drag.
i.e. By having a strip, you’re decreasing the likelihood of using hot / warm exhaust air from that area. That is, it’s a passive airflow pathway control mechanism.
5 Likes