This post has been de-listed
It is no longer included in search results and normal feeds (front page, hot posts, subreddit posts, etc). It remains visible only via the author's post history.
Table content data moved to Nvidia AMD thread. Further updates will be added there.
General Info:
- This table, as well as actual testing and study is WIP (work in progress). And being updated.
- [!] Stock Wattage can be higher and hashrate may be lower vs OC readings. When Overclocking - downvolt core and helps gpu to reach target clocks with that. (works with AMD)
- [!] Software wattage readings (GPU-Z, Hwinfo etc) are often not matching reality with AMD. Try to use hardware meter as much as possible with them. Table represents hardware meters as much as possible (check screenshots where available)
- [!] Consider slight variations depending on your model of a card. Some cards are being OC-ed on stock settings on factory, so they can show higher "default/stock" hashrate vs non-OC models. Always refer to core clock, powerlimit and core voltage when comparing different cards between models and keep in mind that there are many variables to compare (gpu core clock, powerlimits, voltages, cooling that leads to throttling, etc).
- I make it happen with community help. Thank you all, for contributing and providing GPUs for testing and sharing info.
- This is reference for Windows-based systems. Linux-based systems config/results should be nearly identical, but leave couple % difference possibility in mind.
STAK CONFIG:
- Xmr-Stak config values (I - intensity, W - Worksize, T - threads per gpu)
GPU MODEL | STATUS | MIN (hash/Watt) | STOCK (hash/Watt) | MAX (hash/Watt) | REFERENCE LINKS | XMR-STAK CONFIG (I/W/T) | COMMENT |
---|---|---|---|---|---|---|---|
HD 5850 | 200* / 110 | [1] | \Feb. fork data.) | ||||
HD 7950 | 500* / 190 | \Feb. fork data. OC is unknown) | |||||
HD 7970 | 595* / 210 | \Feb. fork data. OC is unknown) | |||||
R7 260x | * | ||||||
R9 280 | 430 | [1] | \Feb. fork data. GPU 972 / Mem 1250) | ||||
R9 390 | 930* / 240 | \Feb. fork data. OC is unknown) | |||||
RX 470 [4] | *** | 733 h/s | 778 h/s | [1] | 1536/16/1 | worksize 16 | |
RX 470 [8] | |||||||
RX 480 [4] | |||||||
RX 480 [8] | |||||||
RX 550 [2] 10 CU | *** | 200/45 | 240/55 | [1] | 480/16/1 | Baffin 10 CU core | |
RX 550 [2] 8 CU | pending | Lexa 10 CU core | |||||
RX 550 [4] | |||||||
RX 560 [4] | *** | 310 / 35 | 340 / 80 | 346 / 40 | [1] | ||
RX 570 [4] | *** | 630 / 70 | 750 / 180 | 780 / 125 | [1] | ||
RX 570 [8] | |||||||
RX 580 [4] | |||||||
RX 580 [8] | *** | 469 / 35 | 820 / 175 | 935 / 145 | [1] | 1728/16/1 | Nitro |
RX 590 | |||||||
Vega 56 | pending | 56 BIOS | |||||
Vega 56 | *** | 905 / 75 | 1550 / 330 | 1740 / 300 | [1] | 1792/16/1 | Reference edition air **flashed to 64 nitro bios**. |
Vega 64 | pending | 56 BIOS | |||||
Vega 64 | *** | 1600 / 290 | 2030 / 365 | [1] | 2016/16/1 | reference ed. | |
RX 5700 | 1626/150** | 1893/206** | [1] [2] [3] | 864/8/2 | \* hardware wattage is unknown yet) | ||
RX 5700XT | *** | 1444 / 90 | 1927@230 | 2134 / 258 | [1] | 960/8/2/2 | \ collaborated with @newtoetn and @cryptonuffie in study) |
Radeon VII | 2110 / 170 | [1] | 1920/16/1 | ||||
R9 Nano | ** | 890/ 170 | [1] | 1864/16/1 |
Status:
- found result reported by a user in community (chats, forums, etc)
* - testing by RyoRU (pending)
** - tested by RyoRU (config tweaking study).
*** - fully tested by RyoRU: (config tweaking study OC testing study). (Can be with owner of a GPU involved in process)
MIN / STOCK / MAX:
- Overclocking core GPU with OverdriveNTool (check reference links for details). Since OC of mem has no sense in CN-GPU memory in most cases is set to 0 or downclocked.
Subreddit
Post Details
- Posted
- 5 years ago
- Reddit URL
- View post on reddit.com
- External URL
- reddit.com/r/XmrStak/com...