Intel Core Ultra 7 265K Custom PC #93
Description
Basic information
- Board URL (official): https://amzn.to/43NTLGk
- Board purchased from: Micro Center
- Board purchase date: November 2024
- Board specs (as tested): 20-core CPU, 64GB DDR5-6000 RAM, ASUS ProArt Z890 Creator Motherboard, AMD AI Pro Ryzen R9700 GPU
- Board price (as tested): $2997.58 (the RAM was $400 less when I bought it, compared to this price)
Linux/system information
# output of `fastfetch`
.... jgeerlingg@asus-proart
.',:clooo: .:looooo:. ----------------------
.;looooooooc .oooooooooo' OS: Ubuntu 25.10 x86_64
.;looooool:,''. :ooooooooooc Kernel: Linux 6.17.0-6-generic
;looool;. 'oooooooooo, Uptime: 24 mins
;clool' .cooooooc. ,, Packages: 1734 (dpkg), 11 (snap)
... ...... .:oo, Shell: bash 5.2.37
.;clol:,. .loooo' Display (GLKVM): 1920x1080 @ 60 Hz in 27" [External]
:ooooooooo, 'ooool DE: GNOME 49.0
'ooooooooooo. loooo. WM: Mutter (Wayland)
'ooooooooool coooo. WM Theme: Yaru
,loooooooc. .loooo. Theme: Yaru [GTK2/3/4]
.,;;;'. ;ooooc Icons: Yaru [GTK2/3/4]
... ,ooool. Font: Ubuntu Sans (11pt) [GTK2/3/4]
.cooooc. ..',,'. .cooo. Cursor: Yaru (24px)
;ooooo:. ;oooooooc. :l. Terminal: Ptyxis 49.1
.coooooc,.. coooooooooo. Terminal Font: Ubuntu Sans Mono (11pt)
.:ooooooolc:. .ooooooooooo' CPU: Intel(R) Core(TM) Ultra 7 265K (20) @ 5.60 GHz
.':loooooo; ,oooooooooc GPU 1: AMD Radeon AI PRO R9700 [Discrete]
..';::c' .;loooo:' GPU 2: Intel Graphics @ 2.00 GHz [Integrated]
Memory: 3.12 GiB / 60.90 GiB (5%)
Swap: 3.91 MiB / 8.00 GiB (0%)
Disk (/): 89.53 GiB / 1.83 TiB (5%) - ext4
Local IP (enp131s0): 10.0.2.224/24
Locale: en_US.UTF-8
# output of `uname -a`
Linux asus-proart 6.17.0-6-generic #6-Ubuntu SMP PREEMPT_DYNAMIC Tue Oct 7 13:34:17 UTC 2025 x86_64 GNU/Linux
Benchmark results
CPU
- Geekbench 6: (3340 single / 21726 multi - https://browser.geekbench.com/v6/cpu/15219330)
- 741.10 Gflops at 273W for 2.71 Gflops/W (geerlingguy/top500-benchmark HPL result)
Power
- Idle power draw (at wall): 34.5 W
- Maximum simulated power draw (
stress-ng --matrix 0): 244.5 W - During Geekbench multicore benchmark: 201.6 W
- During
top500HPL benchmark: 273 W
(When running non-GPU tests, I had the Radeon AI Pro R9700 plugged in and operating, and it draws approximately 10-15W at idle. As such, I've removed 15W of power draw from all the above tests, in lieu of re-testing without a GPU attached. Tests were performed with the monitor off and DisplayPort plug disconnected.)
Disk
KXG80ZNV2T04 KIOXIA 2TB XG8
| Benchmark | Result |
|---|---|
| iozone 4K random read | 82.00 MB/s |
| iozone 4K random write | 404.63 MB/s |
| iozone 1M random read | 4101.48 MB/s |
| iozone 1M random write | 4695.74 MB/s |
| iozone 1M sequential read | 4607.48 MB/s |
| iozone 1M sequential write | 5019.31 MB/s |
Network
iperf3 results:
Built-in 10G Ethernet (Aquantia AQC113)
iperf3 -c $SERVER_IP: TODO Mbpsiperf3 -c $SERVER_IP --reverse: TODO Mbpsiperf3 -c $SERVER_IP --bidir: TODO Mbps up, TODO Mbps down
Built-in 2.5G Ethernet (Intel I226-V)
iperf3 -c $SERVER_IP: 2.36 Gbpsiperf3 -c $SERVER_IP --reverse: 2.35 Gbpsiperf3 -c $SERVER_IP --bidir: 2.33 Gbps up, 2.32 Gbps down
Built-in WiFi 7 (Intel BE200)
iperf3 -c $SERVER_IP: TODO Mbpsiperf3 -c $SERVER_IP --reverse: TODO Mbpsiperf3 -c $SERVER_IP --bidir: TODO Mbps up, TODO Mbps down
GPU
See:
- AMD Radeon AI Pro R9700 general benchmarks in this system: Test AMD Radeon AI Pro R9700 raspberry-pi-pcie-devices#785 (comment)
- AI inference with AMD Radeon AI Pro R9700: Benchmark AMD Radeon AI Pro R9700 32GB ai-benchmarks#35
Memory
tinymembench results:
Click to expand memory benchmark result
tinymembench v0.4.10 (simple benchmark for memory throughput and latency)
==========================================================================
== Memory bandwidth tests ==
== ==
== Note 1: 1MB = 1000000 bytes ==
== Note 2: Results for 'copy' tests show how many bytes can be ==
== copied per second (adding together read and writen ==
== bytes would have provided twice higher numbers) ==
== Note 3: 2-pass copy means that we are using a small temporary buffer ==
== to first fetch data into it, and only then write it to the ==
== destination (source -> L1 cache, L1 cache -> destination) ==
== Note 4: If sample standard deviation exceeds 0.1%, it is shown in ==
== brackets ==
==========================================================================
C copy backwards : 16498.8 MB/s (0.7%)
C copy backwards (32 byte blocks) : 16476.1 MB/s (0.5%)
C copy backwards (64 byte blocks) : 16539.6 MB/s (0.1%)
C copy : 16413.8 MB/s
C copy prefetched (32 bytes step) : 15816.5 MB/s (0.1%)
C copy prefetched (64 bytes step) : 16433.1 MB/s
C 2-pass copy : 12747.5 MB/s (1.0%)
C 2-pass copy prefetched (32 bytes step) : 10635.6 MB/s (0.3%)
C 2-pass copy prefetched (64 bytes step) : 11923.0 MB/s (0.6%)
C fill : 40317.8 MB/s (0.2%)
C fill (shuffle within 16 byte blocks) : 40297.2 MB/s (0.1%)
C fill (shuffle within 32 byte blocks) : 40216.1 MB/s
C fill (shuffle within 64 byte blocks) : 40159.8 MB/s
---
standard memcpy : 37140.8 MB/s (0.3%)
standard memset : 80744.9 MB/s
---
MOVSB copy : 20168.3 MB/s (1.0%)
MOVSD copy : 20158.2 MB/s (0.1%)
SSE2 copy : 18725.3 MB/s (0.3%)
SSE2 nontemporal copy : 28289.2 MB/s (0.2%)
SSE2 copy prefetched (32 bytes step) : 18543.6 MB/s (0.2%)
SSE2 copy prefetched (64 bytes step) : 17370.8 MB/s (0.2%)
SSE2 nontemporal copy prefetched (32 bytes step) : 23816.7 MB/s
SSE2 nontemporal copy prefetched (64 bytes step) : 17776.9 MB/s
SSE2 2-pass copy : 13262.9 MB/s (0.2%)
SSE2 2-pass copy prefetched (32 bytes step) : 11538.4 MB/s (0.4%)
SSE2 2-pass copy prefetched (64 bytes step) : 10292.5 MB/s
SSE2 2-pass nontemporal copy : 3896.6 MB/s
SSE2 fill : 41994.2 MB/s
SSE2 nontemporal fill : 78814.5 MB/s
==========================================================================
== Framebuffer read tests. ==
== ==
== Many ARM devices use a part of the system memory as the framebuffer, ==
== typically mapped as uncached but with write-combining enabled. ==
== Writes to such framebuffers are quite fast, but reads are much ==
== slower and very sensitive to the alignment and the selection of ==
== CPU instructions which are used for accessing memory. ==
== ==
== Many x86 systems allocate the framebuffer in the GPU memory, ==
== accessible for the CPU via a relatively slow PCI-E bus. Moreover, ==
== PCI-E is asymmetric and handles reads a lot worse than writes. ==
== ==
== If uncached framebuffer reads are reasonably fast (at least 100 MB/s ==
== or preferably >300 MB/s), then using the shadow framebuffer layer ==
== is not necessary in Xorg DDX drivers, resulting in a nice overall ==
== performance improvement. For example, the xf86-video-fbturbo DDX ==
== uses this trick. ==
==========================================================================
MOVSD copy (from framebuffer) : 24622.9 MB/s
MOVSD 2-pass copy (from framebuffer) : 25915.8 MB/s
SSE2 copy (from framebuffer) : 25370.8 MB/s
SSE2 2-pass copy (from framebuffer) : 26336.6 MB/s
==========================================================================
== Memory latency test ==
== ==
== Average time is measured for random memory accesses in the buffers ==
== of different sizes. The larger is the buffer, the more significant ==
== are relative contributions of TLB, L1/L2 cache misses and SDRAM ==
== accesses. For extremely large buffer sizes we are expecting to see ==
== page table walk with several requests to SDRAM for almost every ==
== memory access (though 64MiB is not nearly large enough to experience ==
== this effect to its fullest). ==
== ==
== Note 1: All the numbers are representing extra time, which needs to ==
== be added to L1 cache latency. The cycle timings for L1 cache ==
== latency can be usually found in the processor documentation. ==
== Note 2: Dual random read means that we are simultaneously performing ==
== two independent memory accesses at a time. In the case if ==
== the memory subsystem can't handle multiple outstanding ==
== requests, dual random read has the same timings as two ==
== single reads performed one after another. ==
==========================================================================
block size : single random read / dual random read, [MADV_NOHUGEPAGE]
1024 : 0.0 ns / 0.0 ns
2048 : 0.0 ns / 0.0 ns
4096 : 0.0 ns / 0.0 ns
8192 : 0.0 ns / 0.0 ns
16384 : 0.0 ns / 0.0 ns
32768 : 0.0 ns / 0.0 ns
65536 : 0.3 ns / 0.4 ns
131072 : 0.7 ns / 0.9 ns
262144 : 1.2 ns / 1.6 ns
524288 : 1.8 ns / 2.2 ns
1048576 : 2.8 ns / 3.2 ns
2097152 : 3.3 ns / 3.6 ns
4194304 : 6.4 ns / 8.5 ns
8388608 : 11.1 ns / 14.3 ns
16777216 : 14.9 ns / 17.4 ns
33554432 : 27.7 ns / 39.4 ns
67108864 : 64.5 ns / 89.8 ns
block size : single random read / dual random read, [MADV_HUGEPAGE]
1024 : 0.0 ns / 0.0 ns
2048 : 0.0 ns / 0.0 ns
4096 : 0.0 ns / 0.0 ns
8192 : 0.0 ns / 0.0 ns
16384 : 0.0 ns / 0.0 ns
32768 : 0.0 ns / 0.0 ns
65536 : 0.3 ns / 0.4 ns
131072 : 0.7 ns / 0.9 ns
262144 : 1.2 ns / 1.6 ns
524288 : 1.8 ns / 2.2 ns
1048576 : 2.2 ns / 2.3 ns
2097152 : 2.3 ns / 2.4 ns
4194304 : 5.6 ns / 8.0 ns
8388608 : 10.5 ns / 14.3 ns
16777216 : 13.0 ns / 16.3 ns
33554432 : 22.8 ns / 32.5 ns
67108864 : 60.7 ns / 85.8 ns
Core to Core Memory Latency
sbc-bench results
N/A
Phoronix Test Suite
Results from pi-general-benchmark.sh:
- pts/encode-mp3: 4.234 sec
- pts/x264 1080p: 273.47 fps
- pts/x264 4K: 66.27 fps
- pts/phpbench: 1738528
- pts/build-linux-kernel (defconfig): 54.468 sec