How to configure Claymore "strap", "sintensity", "rxboost" and "eres" arguments in Kryptex.
Claymore miner supports special arguments, which helps to fine-tune your video cards and increase hashrate and profitability. Kryptex uses Claymore to mine Ethereum (ETH) on DaggerHashimoto#1 and Expanse (EXP) on DaggerHashimoto#2.
By tweaking the
-rxboost options, you can adjust the timings of your GPUs and increase hashrate by 10-20% depending on a GPU model and memory chips installed. Adding the
-eres 0 parameter will allow you to prolong an ETH mining on the cards where DAG file size start to exceed the video memory size.
⚠ Make sure you've read the instructions thoroughly and got everything clear. Reckless overclocking may cause headaches and a lost profit 💸
Our guide will help you understand how to apply and test the said arguments safely. In case you face any trouble, let us know in the Kryptex community chat — with the help of our team and experienced Kryptex users you'll be able to figure it all out 💪🏻
How to apply extra Claymore arguments in Kryptex?
- Go to Kryptex setting and enable a Pro mode;
- There you will find an "Extra miner args" field;
- Enter arguments separated by space.
⚠ Kryptex disables automatic algorithm switching when extra args are set. You won't benefit from auto-switching to the most profitable algorithm anymore.
-eres 0 — a helping hand for 4GB GPUs
To mine Ethereum, video card must load a special DAG-file into it's memory. The size of the DAG-file is constantly growing, so that less and less video cards are able to mine ETH.
When DAG-file size approaches the size of the video card's VRAM, miner starts to show an unstable behavior — hashrate drops, miner might spontaneously crash. Add the
-eres 0 argument to the extra args — this will free a little bit of memory and will let you mine Ethereum for an epoch or two more.
Kryptex automatically applies
-eres0for you, if you are not using "Extra miner args" or if you are mining in the normal mode.
Claymore args to boost your mining hashrate
-strap — overclock GPU memory timings
Strap is a set of modified memory timings that improves a GPU performance. Claymore miner passes these timings to a video card → video card applies them → performance and profitability increases.
Timing adjustment is a pretty radical overclock method. Your GPU may start to perform weirdly, PC might spontaneously reboot, miner might start crashing and generating bad shares.
Straps don't make permanent modifications to your GPU. An overclock is active only while the miner is on.
Straps are much safer then a custom BIOS firmware. If straps make your rig unstable, simply turn the miner off or reboot your PC. This will roll modifications back to the safe defaults.
To use the
-strap argument you need:
- Kryptex miner v4.4+ or Claymore v15.0+ ;
- AMD RX4xx/RX5xx series (Polaris), Vega or Nvidia 10xx series;
- AMD drivers v18.0+ or Nvidia drivers v400.00+.
How does the
-strap option works?
Claymore miner has a built-in straps database with preset timings for all kinds of different GPUs. All straps are sorted by intensity.
- Polaris cards have a different number of straps, depending on memory type and size.
- Vega cards have 5 presets
- Nvidia cards got 6, but in two groups:
1..3— normal straps and
4..6— low-intensity straps.
Strap "0" does nothing and is used to set a particular strap for each individual GPU (see below).
Enabling on every card:
-strap with strap index you want to apply to your cards.
-strap 1 will apply a strap#1 to all of your cards.
Enabling on particular cards:
Use a comma-separated string of strap indices in the same order as your GPUs appear in the Kryptex app.
-strap 0,1,0,0,0 will apply strap#1 to the second GPU only, and will leave the rest at the stock timings.
-strap 1,0,4 will apply strap#1 to the 1st GPU and strap#4 to the 3rd GPU while leaving the 2nd GPU on stock settings.
More precise tuning
@ symbol to set the memory clock along with the strap. Using this arg is no different from using the MSI Afterburner any other GPU tuner.
-strap 1@2100 will apply strap#1 and a memory clock of 2100MHz to every GPU in your system.
You can also set memory clocks for each of your cards independently:
- strap#1 and 2100MHz clock for the 1st GPU;
- strap#1 and 2300MHz clock for the 2nd GPU;
- strap#2 for the 3rd GPU;
- ...and stock for the 4th GPU.
-sintensity — strap intensity for Nvidia cards.
-sintensity sets the intensity of a particular strap in percents and only works with Nvidia cards.
-strap 1 -sintensity 50 sets strap#1 at 50%.
You can tune per-card intensity by going
-sintensity 10,0,100,30. Zero means stock strap's intensity.
-sintensity is used
Nvidia cards have two kinds of straps: 1…3 are normal ones, while 4…6 are low-intensity straps. Use the latter if the GPU gets too hot or the system becomes unstable.
In some scenarios, even the least intensive straps #1 and #4 might cause problems. Use the
-sintensity option to additionally tune the intensity of a particular strap.
Start with the lowest percent possible:
-strap 4 -sintensity 1 and increase the intensity by 15-20% at the time, until you hit 100%. If that still doesn't help, try it with other straps.
-rxboost — a little extra for AMD cards.
-rxboost 1 to enable the option and get another ~3-5% on AMD Polaris and older Hawaii, Tonga, Tahiti, or Pitcairn GPUs.
If card goes unstable, tune the intensity 2…100%, i.e.
-rxboost 50 for 50%.
Boost also can be applied per-card:
-rxboost 1,0,10,30. Zero means no boost.
What are the perfect settings?
There is no single recipe for success since every GPU and PC is unique. You have to find the perfect parameters by yourself.
Go wild with tests and experiments! Document the results, evaluate them, and draw the conclusions.
Fasten your tester seat belt
- Finding a proper set of args might take some time and effort. Get yourself a pen and paper to record results from what might be your twentieth test run. It's always good to take notes of the process.
- Use Kryptex Pro to track the hash rates, temps, and power consumption. Having every indicator laid out in front of you is quite handy.
- Run your setup for at least 24h straight to reveal any instability.
Proceed with the tests
- Start with the basics — set
-strap 1and check the results. In most cases, that single option is enough to get a significant boost while keeping your rig stable. It's all good? The system runs for 24 hours straight without a wince? Try going with another more powerful straps.
- Does driver stop responding? Does miner keep crashing? That means even
-strap 1is too much for your system.
- For Nvidia try using a low-intensity
-strap 4. If that fails, try lowering intensity with
-sintensity. Start with the lowest percentage possible
-strap 4 -sintensity 1and keep stepping up by 15-20% each time, until you hit 100% or get stable.
- For AMD
-strap 1is already the least intense option. Try lowering core and memory clocks or try using
-rxboost 1without any
- For Nvidia try using a low-intensity
- Even a single failing GPU might crash the whole mining rig. Try testing each card separately to find the weak spot.
- Let's imagine we have an unstable system with 3 GPUs.
- Set the
-strap 1,0,0to test the first GPU only.
- If the first card works fine, go to the next card with
- Keep testing until you find the root of the problem.
Need help setting up Claymore straps?
Reach out to our experienced miners in the Kryptex community chat. Kryptex users and our team be happy to help you out 💪🏻
Or contact us directly!