**This guide assumes you can build a working nheqminer on your system already**
By default, `nheqminer` is configured to build with `XENONCAT` CPU solvers for AVX/AVX2 CPU instruction sets.
If you try to run it on a VPS or other machine which does not have AVX/AVX2 instruction sets on the CPU, **it will result in 0 sols/s hashing speed but report no error while running.**
You may see something like the following:
```
[01:01:55][0x00007fe305412740] Using SSE2: YES
[01:01:55][0x00007fe305412740] Using AVX: NO
[01:01:55][0x00007fe305412740] Using AVX2: NO
[23:50:38][0x00007f4fe6871700] miner#2 | Starting thread #2 (cpu_tromp_STUB)
[23:50:38][0x00007f4fe7873700] miner#0 | Starting thread #0 (cpu_tromp_STUB)
...
[23:51:02][0x00007f4fe9111740] Speed [15 sec]: 0 I/s, 0 Sols/s
...
```
***You need to enable TROMP CPU solver to use regular SSE2 instructions which most VPS should have***
To fix this is simple. First grab `nheqminer` from one of the up-to-date forks:
* `git clone https://github.com/aayanl/nheqminer`
follow the instructions on the github page to get it installed up until the `cmake ../nheqminer` part. (***do not run `cmake ../nheqminer` yet.***)
* https://github.com/aayanl/nheqminer#general-instructions (**skip CUDA installation**)
if you are in the build directory, cd into the `nheqminer` directory
* `cd ../nheqminer`
edit `CMakeLists.txt`
* `nano CMakeLists.txt`
Change the following options like so (set them all to OFF apart from `CPU_TROMP`):

Close and save the file
* `ctrl + x` then `y` then `enter`
Now build the program.
* `cd ../build`
* `cmake ../nheqminer`
* `make -j $(nproc)`
If all is well, the program should compile and you should have a working `nheqminer`
Test that the `TROMP` CPU solver is working by invoking the command:
* `./nheqminer -b 30 -t X` *where X is the number of CPUs you want to test on*
It should now benchmark the speed using `CPU-TROMP-SSE2`. Once finished and confirmed it is working you can go ahead and set it to mine properly on a pool.
*Please note that CPU hashrates are very slow compared to GPU, and without AVX/AVX2 XENONCAT acceleration, the hashrate will be even slower, but this should solve the 0 sols/s issue experienced by various people using non-AVX VPS instances.*
If this guide has helped you, consider a donation:
**ZEN:** `znSTMxvU3AizLV9cAm4iNPT5uLoJ2wbfHy9`
**Private ZEN:** `zcK5A39UwgaufiyUVtVqTXMFQXxxCUCvicvuMxcCE9QrgBMAGW5yCQW9a5zRqwZbYBTCMhTZgyhKH3TMMHq4xwLADQvqrM3`
**BTC:** `1EwGXrmGdiD6Xd8uPnmRufyoWowJ7qpkJ1`