2x GTX 1080ti + 1070 on MD5 /NTLM very very slow

hi folks, im getting with my 2x 1080ti + 1070 on stock very weird speeds with md5 and ntlm. -w 3 and only 1 hash is being cracked, but still very low speeds. max 20917.9 MH/s  for NTLM, 12114.5 MH/s  for MD5 per 1080ti.

System specs:
i7 8700k, 64GB DDR4, MSI Z370 Sli plus, ubuntu 18.04,

below outputs.

hashcat64.bin -m 0 tmp_hash -a3 ?a?a?a?a?a?a?a?a
hashcat (v4.2.1) starting...

* Device #1: WARNING! Kernel exec timeout is not disabled.
             This may cause "CL_OUT_OF_RESOURCES" or related errors.
             To disable the timeout, see: https://hashcat.net/q/timeoutpatch
* Device #2: WARNING! Kernel exec timeout is not disabled.
             This may cause "CL_OUT_OF_RESOURCES" or related errors.
             To disable the timeout, see: https://hashcat.net/q/timeoutpatch
* Device #3: WARNING! Kernel exec timeout is not disabled.
             This may cause "CL_OUT_OF_RESOURCES" or related errors.
             To disable the timeout, see: https://hashcat.net/q/timeoutpatch
OpenCL Platform #1: NVIDIA Corporation
======================================
* Device #1: GeForce GTX 1080 Ti, 2794/11178 MB allocatable, 28MCU
* Device #2: GeForce GTX 1080 Ti, 2794/11178 MB allocatable, 28MCU
* Device #3: GeForce GTX 1070, 2029/8119 MB allocatable, 15MCU

Hashes: 1 digests; 1 unique digests, 1 unique salts
Bitmaps: 16 bits, 65536 entries, 0x0000ffff mask, 262144 bytes, 5/13 rotates

Applicable optimizers:
* Zero-Byte
* Early-Skip
* Not-Salted
* Not-Iterated
* Single-Hash
* Single-Salt
* Brute-Force
* Raw-Hash

Minimum password length supported by kernel: 0
Maximum password length supported by kernel: 256

ATTENTION! Pure (unoptimized) OpenCL kernels selected.
This enables cracking passwords and salts > length 32 but for the price of drastically reduced performance.
If you want to switch to optimized OpenCL kernels, append -O to your commandline.

[s]tatus [p]ause ypass [c]heckpoint [q]uit => s

Session..........: hashcat
Status...........: Running
Hash.Type........: MD5
Hash.Target......: e778f845815adedeXXXXXXXXX
Time.Started.....: Tue Sep 25 23:10:27 2018 (11 secs)
Time.Estimated...: Fri Sep 28 13:48:30 2018 (2 days, 14 hours)
Guess.Mask.......: ?a?a?a?a?a?a?a?a [8]
Guess.Queue......: 1/1 (100.00%)
Speed.Dev.#1.....: 10829.8 MH/s (10.58ms) @ Accel:128 Loops:32 Thr:1024 Vec:1
Speed.Dev.#2.....: 12114.5 MH/s (9.51ms) @ Accel:128 Loops:32 Thr:1024 Vec:1
Speed.Dev.#3.....:  6477.9 MH/s (9.51ms) @ Accel:128 Loops:32 Thr:1024 Vec:1
Speed.Dev.#*.....: 29422.1 MH/s
Recovered........: 0/1 (0.00%) Digests, 0/1 (0.00%) Salts
Progress.........: 316867084288/6634204312890625 (0.00%)
Rejected.........: 0/316867084288 (0.00%)
Restore.Point....: 0/7737809375 (0.00%)
Candidates.#1....: RlizXzus -> ;zurc~de
Candidates.#2....: 3 #erane -> E)c!AJUS
Candidates.#3....: E(*Q-uer -> :WOf*MY1
HWMon.Dev.#1.....: Temp: 61c Fan:100% Util:100% Core:1860MHz Mem:5005MHz Bus:8
HWMon.Dev.#2.....: Temp: 62c Fan:100% Util: 99% Core:1961MHz Mem:5005MHz Bus:8
HWMon.Dev.#3.....: Temp: 50c Fan:100% Util: 98% Core:1797MHz Mem:3802MHz Bus:4


[b]hashcat64.bin -m 1000 tmp_hash -a3 ?a?a?a?a?a?a?a?a -w3

[/b][s]tatus [p]ause ypass [c]heckpoint [q]uit => s

Session..........: hashcat
Status...........: Running
Hash.Type........: NTLM
Hash.Target......: b749e584fc1aaXXXXXXXXXXXXXXXXXXXXX
Time.Started.....: Tue Sep 25 23:15:28 2018 (5 secs)
Time.Estimated...: Thu Sep 27 12:45:43 2018 (1 day, 13 hours)
Guess.Mask.......: ?a?a?a?a?a?a?a?a [8]
Guess.Queue......: 1/1 (100.00%)
Speed.Dev.#1.....: 18193.1 MH/s (50.70ms) @ Accel:128 Loops:256 Thr:1024 Vec:1
Speed.Dev.#2.....: 20917.9 MH/s (44.21ms) @ Accel:128 Loops:256 Thr:1024 Vec:1
Speed.Dev.#3.....: 10025.8 MH/s (49.56ms) @ Accel:256 Loops:128 Thr:1024 Vec:1
Speed.Dev.#*.....: 49140.0 MH/s
Recovered........: 0/1 (0.00%) Digests, 0/1 (0.00%) Salts
Progress.........: 239545090048/6634204312890625 (0.00%)
Rejected.........: 0/239545090048 (0.00%)
Restore.Point....: 0/7737809375 (0.00%)
Candidates.#1....: y*sm$GUS ->  am<yQQU
Candidates.#2....: -li~L(34 -> LynJ~dll
Candidates.#3....: .71erane -> c@#2#xer
HWMon.Dev.#1.....: Temp: 64c Fan: 35% Util:100% Core:1860MHz Mem:5005MHz Bus:8
HWMon.Dev.#2.....: Temp: 66c Fan: 36% Util:100% Core:1961MHz Mem:5005MHz Bus:8
HWMon.Dev.#3.....: Temp: 52c Fan: 32% Util:100% Core:1797MHz Mem:3802MHz Bus:4

here benchmark for ntlm:

Benchmark relevant options:
===========================
* --optimized-kernel-enable

Hashmode: 1000 - NTLM

Speed.Dev.#1.....: 60306.4 MH/s (60.43ms) @ Accel:128 Loops:1024 Thr:1024 Vec:2
Speed.Dev.#2.....: 65369.0 MH/s (55.82ms) @ Accel:128 Loops:1024 Thr:1024 Vec:2
Speed.Dev.#3.....: 32376.5 MH/s (61.48ms) @ Accel:256 Loops:512 Thr:1024 Vec:2
Speed.Dev.#*.....:   158.1 GH/s

Hashmode: 0 - MD5

Speed.Dev.#1.....: 36153.9 MH/s (51.07ms) @ Accel:128 Loops:512 Thr:1024 Vec:4
Speed.Dev.#2.....: 39065.0 MH/s (47.32ms) @ Accel:128 Loops:512 Thr:1024 Vec:4
Speed.Dev.#3.....: 19190.3 MH/s (51.83ms) @ Accel:256 Loops:256 Thr:1024 Vec:4
Speed.Dev.#*.....: 94409.2 MH/s

What could be wrong in my setting?