Home / Intel / Intel Graphics 64EU (Arrow Lake): Performance and Specs

Intel Graphics 64EU (Arrow Lake)

Intel Graphics 64EU (Arrow Lake)
Intel Graphics 64EU (Arrow Lake) graphics card review

Intel Graphics 64EU in Arrow Lake: What Four Xe Cores Deliver

Intel Graphics 64EU arrived in Arrow Lake as a replacement for the significantly weaker integrated graphics in previous generations of desktop Core processors. In its full configuration, the GPU includes four Xe cores and 64 EU, while its frequency in desktop Core Ultra 200 processors reaches 2.0 GHz. In terms of performance, it is no longer in the same class as the previous UHD Graphics, but 64 EU is still not designed for modern demanding games at 1080p.

What Changed After UHD Graphics

The GPU is based on the Xe-LPG architecture. The full configuration contains four Xe cores, 64 EU, 32 TMUs, 16 ROPs, and four ray-tracing units. FP32 performance reaches 2.048 TFLOPS.

DirectX 12 Ultimate, Vulkan 1.3, and OpenCL 3.0 are supported. The media engine provides hardware encoding and decoding for AV1, H.264, and H.265, Intel Quick Sync Video, and resolutions up to 8K.

There is no dedicated video memory-the GPU uses system DDR5. It also has no XMX units. XeSS works through DP4a here, without the XMX acceleration available on discrete Arc GPUs.

Which Processors Use It

64 EU is used in desktop Arrow Lake-S processors and higher-end mobile Arrow Lake-HX processors.

Processor Class Maximum GPU Frequency
Core Ultra 5 245K Desktop 1.9 GHz
Core Ultra 7 265K Desktop 2.0 GHz
Core Ultra 9 285K Desktop 2.0 GHz
Core Ultra 7 265HX Laptop up to 2.0 GHz
Core Ultra 9 275HX Laptop up to 2.0 GHz
Core Ultra 9 285HX Laptop up to 2.0 GHz

The same configuration is used in the Core Ultra 5 250K Plus and Core Ultra 7 270K Plus. Arrow Lake also has Intel Graphics versions with 48 and 32 EU, so GPU performance cannot be determined from the GPU name alone.

Benchmarks in Real Systems

Memory, drivers, and platform settings affect the results, so the specific test configurations are listed below.

Configuration Test Result
Core Ultra 9 285K + MSI MPG Z890 Edge Ti WiFi 3DMark Time Spy Graphics 2192
Core Ultra 7 265K + ASUS TUF Gaming Z890-Pro WiFi 3DMark Night Raid Graphics 26,139
Core Ultra 9 285HX + 32 GB DDR5 3DMark Night Raid Graphics 24,281

The Core Ultra 9 285K scored 2192 points in Time Spy Graphics-the same result listed in the Cputronic database. The graphics also achieved 438 points in Steel Nomad, 20,497 points in OpenCL, and 22,384 in Vulkan.

Night Raid and Time Spy use different workloads, so their scores cannot be compared directly.

Gaming Performance

In Cyberpunk 2077 at 1920x1080, the result is around 20 FPS. To achieve an acceptable frame rate, you will have to lower the resolution or graphics quality, or enable upscaling.

GTA V is considerably less demanding-the result reaches approximately 34.9 FPS at 1080p. Older games and undemanding online titles can be played without a discrete graphics card as long as the settings are not turned up too high.

Game Resolution Result
Cyberpunk 2077 1920x1080 20 FPS
GTA V 1920x1080 34.9 FPS

In demanding games, four Xe cores quickly run into a lack of compute performance. This configuration is already insufficient for 1080p gaming in modern AAA titles.

The Difference Between Core Ultra 5 and Core Ultra 9 Is Small

The Core Ultra 5 245K runs its GPU at up to 1.9 GHz, while the Core Ultra 7 265K and Core Ultra 9 285K reach up to 2.0 GHz. The difference is around 5%.

In practice, the frequency and operating mode of the DDR5 memory can have a greater impact on performance than this difference in GPU frequency. Since there is no dedicated VRAM, the result depends on memory bandwidth and dual-channel operation.

There is no point in paying extra for the Core Ultra 9 for its integrated graphics-both chips use the same four Xe cores and 64 EU.

Where 64 EU Is Enough

A discrete graphics card is not essential for an office or home PC: 64 EU is enough for the desktop, video playback, multiple displays, and light gaming.

In modern AAA games, performance is no longer sufficient for comfortable 1080p gaming without lowering the quality or using upscaling. A discrete graphics card remains necessary here.

Conclusion

With the move to 64 EU, Intel’s integrated desktop graphics is no longer merely a backup display adapter. This GPU can already run undemanding games and provide hardware video acceleration without a separate graphics card.

However, four Xe cores remain entry-level. Demanding games quickly push the GPU beyond a comfortable level of performance. 64 EU eliminates the need for a graphics card in a typical work PC, but it is not a sensible foundation for a gaming system.

GPU Comparison

Compare Graphics 64EU (Arrow Lake) with other GPUs

Please select GPU from the dropdown list.

Basic

Label Name
Intel
Platform
Integrated
Launch Date
October 2024
Former Codename
Arrow Lake
GPU Lithography
TSMC N5P
Model Name
Intel Graphics
Generation
Xe-LPG (Arrow Lake-S)
Base Clock
300 MHz
Boost Clock
1900-2000 MHz
RT Cores
4
TMUs
?
Texture Mapping Units (TMUs) serve as components of the GPU, which are capable of rotating, scaling, and distorting binary images, and then placing them as textures onto any plane of a given 3D model. This process is called texture mapping.
32
Foundry
TSMC
Process Size
5 nm
Architecture
Xe-LPG
Used in CPUs

Memory Specifications

Memory Size
System Shared
Memory Type
DDR5 (System Shared)
Memory Bus
?
The memory bus width refers to the number of bits of data that the video memory can transfer within a single clock cycle. The larger the bus width, the greater the amount of data that can be transmitted instantaneously, making it one of the crucial parameters of video memory. The memory bandwidth is calculated as: Memory Bandwidth = Memory Frequency x Memory Bus Width / 8. Therefore, when the memory frequencies are similar, the memory bus width will determine the size of the memory bandwidth.
System Shared
Memory Clock
System Dependent
Bandwidth
?
Memory bandwidth refers to the data transfer rate between the graphics chip and the video memory. It is measured in bytes per second, and the formula to calculate it is: memory bandwidth = working frequency × memory bus width / 8 bits.
System Dependent

Display and Media

AV1 Encode/Decode
Yes
DisplayPort Extensions
DisplayPort 2.1 UHBR20
H.264 Hardware Encode/Decode
Yes
H.265 HEVC Hardware Encode/Decode
Yes
HDMI Version
HDMI 2.1 FRL
Intel Quick Sync Video
Yes
Max Resolution DP
7680 x 4320 @ 60Hz
Max Resolution eDP
3840 x 2400 @ 120Hz
Max Resolution HDMI
4096 x 2304 @ 60Hz (HDMI 2.1 TMDS); 7680 x 4320 @ 60Hz (HDMI 2.1 FRL)
Max Video Decode Bandwidth
Up to 8K 60 FPS 10-bit HDR
Max Video Encode Bandwidth
Up to 8K 120 FPS 10-bit HDR
Number of Displays Supported
4
Outputs
eDP 1.4b, DisplayPort 2.1 UHBR20, HDMI 2.1 FRL

Theoretical Performance

Pixel Rate
?
Pixel fill rate refers to the number of pixels a graphics processing unit (GPU) can render per second, measured in MPixels/s (million pixels per second) or GPixels/s (billion pixels per second). It is the most commonly used metric to evaluate the pixel processing performance of a graphics card.
30.4-32.0 GPixel/s
Texture Rate
?
Texture fill rate refers to the number of texture map elements (texels) that a GPU can map to pixels in a single second.
60.8-64.0 GTexel/s
FP32 (float)
?
An important metric for measuring GPU performance is floating-point computing capability. Single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks, while double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable.
2.048 TFLOPS

AI Features

AI Software Frameworks Supported by GPU
OpenVINO, WindowsML, DirectML, ONNX RT, WebGPU
GPU Peak TOPS (Int8)
8 TOPS
Intel Deep Learning Boost on GPU
Yes

Miscellaneous

Shading Units
?
The most fundamental processing unit is the Streaming Processor (SP), where specific instructions and tasks are executed. GPUs perform parallel computing, which means multiple SPs work simultaneously to process tasks.
512
L2 Cache
4 MB
Vulkan Version
?
Vulkan is a cross-platform graphics and compute API by Khronos Group, offering high performance and low CPU overhead. It lets developers control the GPU directly, reduces rendering overhead, and supports multi-threading and multi-core processors.
1.3
OpenCL Version
3.0
OpenGL
4.5
CUDA
No
DirectX
12 Ultimate (12_2)
ROPs
?
The Raster Operations Pipeline (ROPs) is primarily responsible for handling lighting and reflection calculations in games, as well as managing effects like anti-aliasing (AA), high resolution, smoke, and fire. The more demanding the anti-aliasing and lighting effects in a game, the higher the performance requirements for the ROPs; otherwise, it may result in a sharp drop in frame rate.
16
Shader Model
6.8

Benchmarks

Cyberpunk 2077 1080p
Score
20 fps
GTA 5 1080p
Score
34.9 fps
FP32 (float)
Score
2.048 TFLOPS
3DMark Steel Nomad
Score
438
3DMark Time Spy
Score
2192
Vulkan
Score
22384
OpenCL
Score
20497

Compared to Other GPU

Cyberpunk 2077 1080p / fps
Radeon RX 6700 XT
75 +275%
Radeon RX 6700
60 +200%
GeForce RTX 2060 12 GB
47 +135%
Arc 140T
23 +15%
Graphics 64EU (Arrow Lake)
20
GTA 5 1080p / fps
Radeon RX 7800 XT
186 +433%
GeForce GTX 1050 Ti
172 +392.8%
GeForce GTX 1660 Ti
151 +332.7%
Radeon RX 480
108 +209.5%
Graphics 64EU (Arrow Lake)
34.9
FP32 (float) / TFLOPS
GeForce GTX 775M Mac Edition
2.185 +6.7%
FirePro D300
2.132 +4.1%
Graphics 64EU (Arrow Lake)
2.048
Radeon RX Vega 11 Mobile
2.01 -1.9%
FirePro V7800
1.976 -3.5%
3DMark Time Spy
Radeon R9 390X
4330 +97.5%
GeForce GTX 1650 Ti Max Q
3227 +47.2%
Graphics 64EU (Arrow Lake)
2192
Radeon Vega 7
1420 -35.2%
Radeon Graphics 2-Core
715 -67.4%
3DMark Steel Nomad
GeForce RTX 3050 Ti Mobile
467 +6.6%
Radeon 860M
454.3 +3.7%
Graphics 64EU (Arrow Lake)
438
GeForce RTX 3050 Mobile
419 -4.3%
GeForce RTX 2050 Mobile
394 -10%
Vulkan
GeForce GTX 1660 Ti
61425 +174.4%
GeForce GTX 980
38993 +74.2%
Graphics 64EU (Arrow Lake)
22384
Radeon Pro WX 2100
10891 -51.3%
GeForce 940M
5522 -75.3%
OpenCL
Arc B390
54698 +166.9%
Radeon RX 580 2048SP
34827 +69.9%
Graphics 64EU (Arrow Lake)
20497
Quadro K4200
12186 -40.5%
Radeon Vega 5
8633 -57.9%

Related GPU Comparisons

Intel Arc B370
Intel Arc B370
VS
Intel Graphics 64EU (Arrow Lake)
Intel Graphics 64EU (Arrow Lake)