Home Artists Posts Import Register

Content

At the moment HYDRA already contains 3 presets of voltages.
I want to add an automatic best volt search function that will take into account your cooling system. What controls would you like to see for this feature? A temperature threshold setting? Maybe something else ? Or do you have any other ideas?

Comments

Beech Horn

Some way to ensure the voltage isn’t too low for the machine to reboot by itself on Zen 3 and presumably some way to track PBO settings against voltage (so upping the PBO limits as the voltage drop gives headroom let’s you see prior results with different PBO options). Also if we can make a note of the ambient temperature of each run (or select an existing temp sensor for ambient).

Anonymous

I want to see the guide line of the performance related to the freq., voltages, and the temp. on the hydra. If it is possible, anyone who want something could know what they have to prepare or pay.

Anonymous

Maybe a setting for preheating the water for watercooled systems. With a custom it can take quite a while till the water is saturated and I want my system stable at those times as well. Currently I let prime95 run for 30min before I start the analystics.

Anonymous

Some way to poll for liquid temp would be intersting.. I'm personally not verse enough to really know the best way to use that data but I do think it would be valuable to have. The trouble of course would be polling from AiOs.. some give this data, some don't and not all use open api/plugins.. on a custom system this would be easier as many of those either have such built in a block or the pump.. or can be added with a temprature-fitting. Simularly I've heard some who use temp-probes on either side of their Raids to get a before/after and use that somehow for their fans

Anonymous

I'm an intermediate user when it comes to knowing what I'm talking about - but a complete rookie when it comes to overclocking - that's why I'm a Patreon subscriber and so thankful for your work Yuri! I have a Deepcool RGB Castle 360mm. In BIOS I leave the pump on full speed the whole time and recently replaced the stock fans with Lian Li uni fans. When I look at CPU temps in Ryzen Master or CTR the temps seem to ramp up to 60-70c under load in no time, then make their way back to 30c within 30 seconds of the load stopping. When I put my hand in front of the fans there doesn't seem to be much heat emanating unless there has been intense load for 5+ mins, so I really don't understand the relationship between the CPU temp and the water temp inside the loop. Whenever I've had instability issues with CTR it has usually been under low loads, and I'm generally only running the uni fans at about 50% speed when not doing anything intense. I have no idea how many users are as clueless as me about the above issues, but I imagine quite a few of us, or we wouldn't be relying on your amazing software! But these are the questions I face weekly. If there's some way to report to the user what's going on and what recommendations are made based on crashes or performance issues within the dashboard, that would really help people like me out.

Anonymous

like manualøly putting in amb temp?if theres no such sensor on mb,oc.i run 5600x in aorus x570 so i have 2 temp sensors i can connect to it.i currenly use 1 for controlling an extra fan for more intake to gpu.ur idea is really good.

Anonymous

reading some of the comments her already,think i can wrap them up to easier describe what some,me included,seems to want;maybe adding an in-app feature to be able to directly read some of the temp sensors from the mobo.that would be a really QoL feature for many i think.i have 2 sensors i can connect to mobo,they both have 30-40 cm wires so i can monitor everything in my system.but i currently have no way to read then after leaving the bios(that i know of).i have no idea how complicated this would be to achieve but it would be awesome!

Anonymous

Setting it by cooling solution is a great idea - will just need some kind of comprehensive test at each PX state - I find especially with the 5950X on high loads, the default voltages are manageable 1-6 threads, but from 7-12 it's almost worth skipping straight to P2 because it gets suuuper hot.

Anonymous

Max voltage for overclock before the performance starts to deteriorate

Anonymous

Will Hydra include a new DRAM Calculator or is that still planed as a separate release?

Anonymous

I assume you'll need to understand from the user what kind of cooling system is installed, and based on that set certain parameters in Hydra. Not sure what those parameters would be, but certainly the ability to read some ambient sensors would be great, which I assume you'll need to consider for any diagnostic step anyway for calculating the proposed default values/profiles. Interested to hear your thoughts on this point.

Anonymous

It takes time for heat to get through the CPU block and then again it takes quite some time to get the temp of the whole water in your loop warm as only the small amount flowing over your CPU gets heated up. For just short load spikes it is quite normal that you don't see an effect in the cooling. This is also the reason why the order of components in a loop doesn't matter and you don't need to let the GPU water flow through a radiator first before going to the CPU. Additionally for smaller processes like 7nm we get quite some heat spots as a lot of stuff is happening in a very small area. For thermals this means: small area -> harder to get the heat away from there. This is why chips with small structures can be warmer than ones with bigger structures. Hope this helps a bit.

Anonymous

A target max temperature would be useful. Some way to search for the best balance between clock speed, and voltage that will result in a max temp with the temp being the variable input by the user.

Anonymous

Maybe settings that would gear towards what kind of cooling. 1:Air 2:AIO 3:Custom loop I've tweaked my system to get the max out. You have to watch temps vs VID vs MHZ for max stability. mine i can run P1 4.5ghz 1.250v prime95 small fttps about 80c. Theres a point just a lil bit of voltage creates too much heat, you gota balance the 3

Tom DePaolo

It would be nice if you could change profiles from the taskbar. Also it would be useful to have different core profiles. As in the ability to change all profiles at once, so if you have settings that work good you can save those and create a new profile if you want to test things.

Tom DePaolo

Another feature that would be pretty awesome would be the ability to detect what bios settings are being used and give recommendations for better performance. Especially for PBO+ and CO. Since tweaking bios settings can take a lot of trial and error any advice would be great.

Anonymous

It would be a amazing to have task bar icons for temp readings. A step further would be fan curve optimization, but that's a.. dream obviously lol. In regards to CPU optimization, aggressive undervolting profile creation would be awesome. For example, diag runs well and I get golden CPU on 5950x on my SFFPC build, but gaming loads = 85-90 degrees, which I don't want. I would rather have it be around 75-80 max, so I have to tinker a lot with the CTR settings. If this were automated somehow, it would help a lot of beginner users. I don't need max OC performance, but I need max performance given the limited thermal headroom that I have. Which defeats the purpose of having a 5950x, but I may upgrade to a full custom loop in the future. Thank you!