Jump to content

Coddeh

Member
  • Posts

    10
  • Joined

  • Last visited

Awards

This user doesn't have any awards

Coddeh's Achievements

  1. So I've done a fair bit more testing and have come to the conclusion that the Zotac GTX970 blower card is one that is locked to 1212mV and no amount of voltage bios tweaking will push it past that. The stock voltage table goes up to 1281.3mV but wen ever it used to boost the voltage would drop to 1870mV or 1620mV. After changing some voltage table values to the max 1212mV when the card boosts it and there are no perfcaps it will stick to that voltage. The card tops out around the 1550mhz core which is about right for low end GTX970s. Now what I want to know is how I get around the PWR perfcap? I've adjusted the power table values to the theoretical max of 225w for TDP and POWER limit (PCI-E and 2x 6pin) and adjusted the PCIE slot and power to 75w but I'm still hitting power limit it seems when running super position or 3dmark which is dropping my core clock speed. Judging by the GPU-Z screen shot it looks like my board power draw and pcie-slot power isn't anywhere near theoretical max. Is that likely to be a misreading or is something wrong? GPU-Z does show that the graphics card is running at 8x rather then 16x but that shouldn't effect power draw should it?
  2. So after some more testing and bios flashing. It appears that the monitoring software I'm using (GPU-Z and msi afterburner) is unable to read past 1212mV which is why it looked like my voltage table bios mods weren't doing anything. It requires some more testing to make sure but it seems that way because after a bit of overclocking I managed to hit 1529.8mhz on core and 1980mhz on ram (not fully finished ram clocking though) which seems unlikely at 1212mV. The voltage table is set to 1287mV max, I've also noticed that the voltage slider on msi afterburner does nothing for stability so probably doesn't work. Also the perfcap reason in gpuz is mostly on pwr but it was on pwr when the core frequency was at 1250 and adding another 250mhz to it didn't seem to matter while giving a substantial score increase on the superposition benchmark. I've now just managed to pip what was the highest scoring GTX970 on there at 1080p high. Got some more testing to do but ill probably dial clocks and volts back a little as the gpu vrm coil whine has gotten much louder. They were silent and now they are audible under load. Max temp of the gpu after several superposition benchmark runs was 51c. The accelero xtreme IV is a beast.
  3. xAcid, thanks for your help. I've done as you have said and the card is boosting properly now. I just need to figure out why volts are stuck at 1212mV. I have a lot of headroom when it comes to tdp and temperature so I adjusted the voltage table to allow up to 1250mV as per toolbox11's overclocking guide but still a no go.
  4. Hi Guys, Let me start by saying I have a GTX970 and I know its an old card but it has served me well so far. I've been through Kets Definitive Maxwell II (2) vBIOS Modding guide and toolbox11's overclocking guide as well but I'm at a bit of a loss. The specific card I have is a Zotac GTX970 blower which has remained stock ever since I got it mainly because of the awful cooler design and how loud it gets. It's the only graphics card I've never overclocked. I generally push all my components to the max when I can. Recently the blower fan was annoying me at idle so I managed to pickup an arctic accelero xtreme IV for £20. My temps have dropped from 80c stock clocks at load to 49c and at idle its super quiet. I'm a bit of a tinkerer and bios modding doesn't scare me so I've had a go but I'm not really getting any different results. All I've changed are the power limits allowing the theoretical max hoping the card will maybe boost higher which it really doesn't The power supply is a hx520 with 6+2 cables but the card only takes two 6pin allowing for a theoretical max of 225w (pci-e slot plus 2 6 pins). GPU-Z seems to suggest my limiting factor is pwr so am I essentially screwed because of the max power draw when it comes to overclocking? When I put in a manual overclock on the core the card drops down to 1000mV which I can't explain either. I've seen in the guides to mod the voltage table around the clk65 mark but tbh my card doesn't seem to get past the clk43 mark. I've added images of maxwell bios tweaker and monitoring software at the stock state, increased power limit state and manual core clock overclock state. If anyone has any ideas that would be really helpful! Specs: i7 4790k 4.5gh@1.25v or 4.7@1.32v when gaming asrock z97e ac itx 16gb 1600mhz ram zotac gtx 970 blower with xtreme IV samsung 860 ssd corsair hx520
  5. I've checked the link and I would still say that the ram you bought should be fine. It is unlikely that the ram would flat out not work on your motherboard. The only possibility I can see is not being able to overclock your ram as high but that's still dubious. C
  6. In all likelihood it will work just fine, I wouldn't worry about it. While I can't say for sure I imagine it's very similar if not the same ram just renamed. In my experience building PC's I've never had issues with incompatible ram between Intel and AMD. While AMD did have a spell of ram issues at the beginning of ryzen I believe most of that is resolved now. I'm sure if you use it everything will work as it should. C
  7. So as an update I've been running the animation render for 3+ hours and have maxed out at 107c on the vrms. Pretty damn toasty but at least it doesn't thermal throttle like before. As I very rarely do an animation render this seems ok considering the system. If I am to do a render of the work for a client it's almost always a still which doesn't get as hot. Thinwalrus I have two 60mm fans above the vrms right now. As I said in the opening post I used to have a noctua a9x14 just for the vrms which I switch to the two smaller gelid fans for more focused air flow. It's much louder now but the vrms are under control. Luckily I don't have to deal with the sound much as my everyday workload doesn't stress the system. So to summarise vrms we pinging to 120c when rendering in 20 to 30 mins. I swapped the stock thermal pad to the Gelid gp extreme 0.5mm. Rending for 3+ hours and max temp on vrms is 107c. Thanks for your help chaps. C
  8. Hi GeoSquigg, I have already seen that product but I'm not keen on electrically conductive pads for this area on the motherboard. I think the best pads that are not electrically conductive are the Fuji ones at 17 W/mk I've installed the Gelid pads now and am testing keyshot animation render. So far it's peaked at 106c and now down to 104c. It's only been going about 20mins though so I will need to leave it for a bit longer I think. I think this mod has most likely just bought me some extra render time but the vrms will eventually hit 120c. It's late now but I'll do further testing tomorrow. Cheers.
  9. So I've decided to go down the route of replacing the thermal pads on the VRM heatsink to something better. I've ordered the Gelid solutions gp extreme in hope that would increase the thernal efficiancy of the stock cooler. I think the stock thermal pads have a conductivity of 3 W/mk and the Gelid pad is rated at 12 W/mk.
  10. Disclaimer: Cooling Issues with fairly powerful parts in the Dan case a4 Hi guys, I've built a PC to use at work. I mainly 3D sculpt using ZBrush and use other programs such as handbrake for encoding videos and Keyshot for some renders. It works brilliantly however I've been having issues with the VRM's over heating (120C and then thermal throttling) when rendring in keyshot for about 20 - 30 mins. I thought I had fixed the issue when I replaced the noctua A9x14 that had been placed over the VRMs for cooling with two smaller Gelid silent 6 60mm fans however I recently rendered out an animation scene for well over an hour and the VRMS hit 120C again. I was thinking of replacing the stock vrm cooler with: http://www.enzotechnology.com/mos-c1.htm or trimming this one down to size. http://www.enzotechnology.com/mst_dfi_dk.htm The stock VRM heatsink doesn't seem terrible to me in terms of surface area though. My system is as follows: Dan case a4 v2 Ryzen 2700x @ 3.9 1.225v set to static voltage and clock speed gigabyte AB350N wifi gtx 960 corsair h60i with noctua a12x15 Corsair SF600 32gb cosair vengence 3200mhz samsung 960pro nvme Everything else runs fine and temperatures are under control however every now and then when I produce quick keyshot renders it would be nice to not have thermal throttling after 20 - 30 mins. Cheers
×