# Half-precision floating-point vectors Perform operations on vectors that contain half-precision floating-point elements. ## Discussion The simd library provides many functions in this group, in three variants: - The default variant—for example, [`simd_rsqrt(_:)`](/documentation/simd/simd_rsqrt(_:)-i5ka) - The precise variant—for example, [`simd_precise_rsqrt(_:)`](/documentation/simd/simd_precise_rsqrt(_:)-136bw) - The fast variant—for example, [`simd_fast_rsqrt(_:)`](/documentation/simd/simd_fast_rsqrt(_:)-8p6tk) The precise variants are accurate to a few units in the last place (ULPs). The fast variants provide greater speed but may have as little as 8 bits of accuracy for half-precision functions. The compiler flag defines the behavior of the default variant. Ordinarily, the compiler resolves the default variants to their precise counterparts. Set the `-ffast-math` compiler flag to specify that the default variants of the functions resolve to the fast variants. ## Topics ### Vector data types [`typealias simd_half1`](/documentation/simd/simd_half1) A vector of one 16-bit floating-point element. [`typealias simd_half2`](/documentation/simd/simd_half2) A vector of two 16-bit floating-point elements. [`typealias simd_half3`](/documentation/simd/simd_half3) A vector of three 16-bit floating-point elements. [`typealias simd_half4`](/documentation/simd/simd_half4) A vector of four 16-bit floating-point elements. [`typealias simd_half8`](/documentation/simd/simd_half8) A vector of eight 16-bit floating-point elements. [`typealias simd_half16`](/documentation/simd/simd_half16) A vector of sixteen 16-bit floating-point elements. [`typealias simd_half32`](/documentation/simd/simd_half32) A vector of thirty-two 16-bit floating-point elements. --- Copyright © 2026 Apple Inc. All rights reserved. | [Terms of Use](https://www.apple.com/legal/internet-services/terms/site.html) | [Privacy Policy](https://www.apple.com/privacy/privacy-policy)